Integrating machine-learning based image analysis with sample deposition and staining methods and systems

Abstract

Methods and systems directed to benchtop instruments for sample deposition and staining with machine-learning based image analysis are described. An example system includes an apparatus with a slide processing module, positioned at an upper front position of the apparatus, which has a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample. The system further includes a clamping module positioned at a lower front of the apparatus, an auxiliary systems module positioned at a rear of the apparatus, and an optical imaging and analysis system configured to analyze the cellular sample either on-site or remotely. An example method for depositing, staining and analyzing a cellular sample uses the above-described apparatus.

Description

INTEGRATING MACHINE-LEARNING BASED IMAGE ANALYSIS WITH SAMPLE DEPOSITION AND STAINING METHODS AND SYSTEMSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to both U.S. Provisional Patent Application No. 63 / 547,622 filed November 7, 2023, and U.S. Provisional Patent Application No. 63 / 686,100 filed August 22, 2024, the entire contents of which are incorporated herein by reference.TECHNICAL FIELD

[0002] This document generally relates to laboratory procedures, and more specifically, to sample deposition and staining with machine-learning (ML)-based image analysis for laboratory cytology procedures.BACKGROUND

[0003] Cytology (also known as cytopathology) involves examining cells from bodily tissues or fluids to determine a diagnosis. Cytology techniques have developed to be minimally invasive and have revolutionized the practice of medicine. The ability to quickly obtain high quality samples with little discomfort has generally made such procedures more acceptable. More recently, sample deposition and staining devices and instruments include integrated imaging and analysis modules that are designed to improve laboratory practices whilst ultimately delivering better patient care.SUMMARY

[0004] Embodiments of the disclosed technology are directed to benchtop instruments for cellular sample deposition and staining that are functionally integrated with machine-learning (ML) based imaging and analysis. The disclosed embodiments provide a rapid, on-site, digital diagnostic platform that can enhance diagnosis.

[0005] In an example aspect, a system for analyzing a cellular sample on a substrate includes a microscope configured to generate a macroscopic view of a surface of the substrate at a first magnification level of a set of magnification levels, a slide scanner, coupled to the microscope, configured to generate a digitized view of the surface of the substrate at each of the set of magnification levels, and one or more processors, coupled to the microscope and slide scanner, configured to identifycoordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection. In this system, the slide scanner is further configured to generate, at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view, and the one or more processors is further configured to identify one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection. Additionally, the slide scanner is further configured to generate, at a third magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile, and the one or more processors is further configured to identify, based on the second digitized representation, at least one cell type within the samples of interest.

[0006] In another example aspect, an apparatus for depositing, staining, and analyzing a cellular sample includes a slide processing module positioned at an upper front position of the apparatus, an auxiliary systems module positioned at a rear of the apparatus, and an optical imaging and analysis system. In this apparatus, the slide processing module includes a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample, and a clamping module positioned at a lower front of the apparatus; the auxiliary systems module includes a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol-based fixative bottle, and an electronics subsystem configured to execute, based on the input from the user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on a surface of a substrate; and the optical imaging and analysis system, which is configured to analyze the stained cellular sample, includes a microscopy subsystem configured to generate a digitized representation of the surface of the substrate at one or more magnification levels, and one or more processors configured to identify one or more portions of the digitized representation that include the stained cellular sample.

[0007] In yet another example aspect, a method of analyzing a cellular sample on a substrate includes generating, using a microscope at a first magnification level, a macroscopic view of a surface of the substrate, and identifying, using one or more processors, coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection. Themethod further includes generating, using a slide scanner at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view, and identifying, using the one or more processors, one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection. Finally, the method includes generating, using the slide scanner at a third magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile, and identifying, based on the second digitized representation, at least one cell type within the samples of interest.

[0008] In yet another example aspect, methods for depositing and staining a sample, followed by ML-based image analysis, which use the apparatus described above are provided.

[0009] In yet another example aspect, one or more operations of the described method can be embodied in the form of an apparatus that includes a processor and a memory coupled to the processor.

[0010] In yet another example aspect, one or more operations of the described method can be embodied in the form of processor-executable instructions and stored on a computer-readable program medium.

[0011] The above and other aspects and their implementations are described in greater detail in the drawings, the descriptions, and the claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIGS. 1 and 2 show examples of systems that functionally integrate a benchtop instrument for sample deposition and staining with machine-learning (ML)- based imaging and analysis.

[0013] FIG. 3 shows an example of a system that both functionally and physically integrates ML-based imaging and analysis into a benchtop instrument for sample deposition and staining.

[0014] FIG. 4 shows examples of different workflows using a benchtop instrument for sample deposition and staining with ML-based imaging and analysis.

[0015] FIG. 5 shows an external view of an example benchtop instrument.

[0016] FIG. 6 shows a side open-view of the benchtop instrument.

[0017] FIGS. 7A and 7B show open-views of both sides of an example auxiliary system of the benchtop instrument.

[0018] FIG. 8 shows an example of a chassis of the auxiliary system.

[0019] FIG. 9 shows an example of a chassis spine that separates the “fluid” side of the benchtop instrument from its “dry” side.

[0020] FIGS. 10A and 10B show an example of a false wall, which provides a neat appearance to the bottle cavity, in the benchtop instrument.

[0021] FIGS. 11 A-1 1 C show examples of the pneumatic system used in the benchtop instrument for spray deposition.

[0022] FIG. 12 shows an example of the drip tray in the benchtop instrument.

[0023] FIGS. 13A and 13B show an example of a slide processing mechanism, which is further detailed in FIGS. 14A-31 B.

[0024] FIGS. 14A-14G show an example of a SIP dock.

[0025] FIG. 15 shows an example of the geometry of the air nozzle.

[0026] FIG. 16A shows an example operation of the SIP side spring.

[0027] FIGS. 16B and 16C show an example operation of the SIP switch.

[0028] FIG. 17 shows an example of the SIP manifold.

[0029] FIGS. 18A-18D show an example of a removable cassette.

[0030] FIGS. 19A and 19B show an example of the slide beds.

[0031] FIGS. 20A shows an example of a slide bed.

[0032] FIG. 20B shows an example of slide switches.

[0033] FIG. 21 shows an example of a slide spring bed.

[0034] FIG. 22 shows an example of an ethanol fixative system.

[0035] FIG. 23 shows an example of an inlet fitting.

[0036] FIG. 24A and 24B show examples of an ethanol spray insert.

[0037] FIG. 25 shows an example of a buffer system.

[0038] FIG. 26 shows an example of a buffer nozzle located above the SIP well.

[0039] FIGS. 27A and 27B show an example operation of an air curtain.

[0040] FIGS. 28A and 28B show an example of a reagent manifold assembly.

[0041] FIGS. 29A and 29B show an example of hood switches.

[0042] FIGS. 30A-30C show an example of a scavenger port.

[0043] FIGS. 31 A and 31 B show an example of reagent paths.

[0044] FIG. 32A shows an example of a clamping mechanism.

[0045] FIG. 32B shows an example handle of the clamping mechanism.

[0046] FIG. 33 shows an example solenoid latch of the clamping mechanism.

[0047] FIG. 34 shows examples of covers and doors of the benchtop instrument.

[0048] FIGS. 35A and 35B show an example of a door providing access to the inner components and circuitry of the benchtop instrument.

[0049] FIGS. 36A and 36B show examples of the front and read handles for the benchtop instrument.

[0050] FIG. 37 shows an example of an ML-based classification system used with the benchtop instruments described herein.

[0051] FIGS. 38A-38C show an example of a stained slide, the slide after being processed by the Specimen Region Identification (SRI) algorithm, and the slide after being processing by the Relevant Tile Identification (RTI) algorithm, respectively.

[0052] FIGS. 39A-39C show another example of a stained slide, the slide after being processed by the SRI algorithm, and the slide after being processing by the RTI algorithm, respectively.

[0053] FIGS. 40A and 40B show screenshots of an example user interface for the imaging and analysis of a cellular sample.

[0054] FIG. 41 shows a flowchart for an example method of analyzing a cellular sample on a substrate.DETAILED DESCRIPTION

[0055] To make the purposes, technical solutions and advantages of this disclosure more apparent, various embodiments are described in detail below with reference to the drawings. Unless otherwise noted, embodiments and features in embodiments of the present document may be combined with each other.

[0056] Section headings are used in the present document to improve readability of the description and do not in any way limit the discussion or the embodiments to the respective sections only.

[0057] 1 Introduction

[0058] Biological tissue samples are collected from patients for microscopic and molecular diagnostic analysis for clinical, diagnostic and research applications. These samples are collected in a variety of laboratory, medical clinic and other health-care ormedical research settings. For example, cells / tissue can be collected from a patient using a collection device, such as a brush, swab or cutting tool for biopsies.

[0059] Rapid on-site evaluation (ROSE) is a cytopathologic diagnostic adequacy assessment of individual biopsy passes performed during a biopsy procedure in order to optimize the procedure itself and inform subsequent patient management. One of the goals of ROSE is to optimize the quantity of diagnostic material procured during a biopsy procedure (e.g., a cellular sample) including triage in anticipation of any ancillary studies (e.g., flow cytometry, molecular testing, etc.). In an example, ROSE specimens consist of aspirate smears from fine needle aspiration (FNA) biopsies or touch preparation slides from needle core (NC) biopsies.

[0060] The subsequent histological or cytological analysis of these cell samples depends greatly on the quality of prepared specimens. Inadequate preparation of specimens can result in inaccurate data that causes an error in interpretation of results and misdiagnosis. In addition, complicated, non-automated systems lead to backlogs, delayed diagnoses, unintended types of artifacts that are due to the complexity of procedures, and intolerable run-to-run variability.

[0061] Current ROSE implementations support remote viewing of stained specimen slides, but there still remain significant drawbacks. For example, in existing systems, pathologists use their keyboard or mouse to control X, Y, and Z actuation stages, as well as to operate other microscope controls. The major challenge with existing remotely-operable microscopes is poor usability for pathologists, who have to physically move their mouse or use multiple keyboard buttons to perform the corresponding movement in the remotely-operable microscope. This procedure requires ultra low-latency communication capabilities and a learning curve that takes a significant amount of time, which has consequently led to poor uptake of remotely-operable microscopes by pathologists.

[0062] Furthermore, remotely-operable microscopes that examine the specimen slide are used by cytologists in non-standardized ways. That is, a pathologist remotely examining the specimen slide in real time is likely following his or her own methods or techniques to determine whether the cellular sample includes cells of a certain size and / or type (e.g., based on past experience or pedagogical methods). The alternative remote-viewing approach is a more systematic procedure for examining the slide, which can be implemented using pathology scanners and ML-based algorithms.

[0063] Digital slide (or pathology) scanners work in conjunction with the remote- operable microscope to allow for the complete scanning of the specimen slide using automated microscope systems. These microscopes quickly digitize the entire specimen slide or a region within a specimen slide. The digitized images are then uploaded to a cloud for it to be accessed by the pathologist / cytopathologist at a later time. Alternatively, these images can also be quickly emailed or shared via other means with the clinicians. Since a remote user only needs access to the digital representation of the specimen slide (or a portion of thereof), there is no longer any need for the remote user to engage with the mouse- or keyboard-interfaced actuator controls.

[0064] However, in existing systems, the amount of time it takes the digital slide scanner to scan a single specimen slide is a major drawback. Typically, it takes about 20 minutes to scan an area of 25mm x 50mm with a 40X objective lens. Although it takes less time (about 5-7 minutes) to scan the same slide area with a 4X or a 10X objective lens, these objective lenses do not provide the necessary resolution to obtain the definite diagnosis. Even if the pathologist were to wait for the higher-resolution scan to be available, his or her own techniques to examine the slide would then be employed, which might not necessarily be the optimal approach for reading the specimen slide.

[0065] Thus, there remains a need for a system that produces reliably consistent stained slide specimens that can be remotely-examined by a cytologist in a timely and systematic manner. Embodiments of the disclosed technology address these drawbacks by providing a benchtop instrument that can reliably produce consistent specimen slides and an ML-based software system that combines the advantages of the remote microscope and the digital slide scanner while resolving the challenges associated with each of the component technologies.

[0066] In some examples, the ML-based software can systematically scan the specimen slide at a lower resolution (e.g., 5X or 10X), while allowing the (remote) cytologist to examine the specimen slide in real time. The lower resolution scans are analyzed, and smaller portions of the scan are identified if they are determined to include any cells of interest. Once identified, these portions can be scanned at a higher resolution (e.g., 40X), and provided to the cytologist. Thus, a systematic scan of the specimen slide can be performed, which can lead to more rapid identification of any cells of interest; first by the ML-based algorithms, and then by the cytologist.

[0067] FIGS. 1 -3 show examples of different embodiments that integrate benchtop instruments for cellular sample deposition and staining with ML-based imaging and analysis. FIGS. 1 and 2 show examples of systems that functionally integrate a benchtop instrument for sample deposition and staining with ML-based imaging and analysis. In these examples, the slide is physically taken from the sample preparation device and loaded into the microscope.

[0068] In some embodiments, the sample preparation device contains a slide holder (e.g., as shown in US D862,727) that can be easily taken out from the sample preparation device and inserted into the microscope.

[0069] In some embodiments, the microscope shown in FIGS. 1 and 2 include the following features and aspects:

[0070] - a display head that shows the field of view of the slide;

[0071] - a connector that enables it to be connected to a computer and for the slide to be viewed using a monitor;

[0072] - a remote connection such that the slides can be visualized from a computer by a clinician who is not physically present in the procedure suite;

[0073] - software with features that enable its easy integration into electronic medical record (EMR) systems or any clinical lab software;

[0074] - additional imaging elements to visualize structures between 20-400 nm within cells; and / or

[0075] - features and / or components that can be used to obtain bright field, dark field, phase contrast or fluorescent images of the cells present on the glass slide.

[0076] FIG. 3 shows an example of a system that both functionally and physically integrates ML-based imaging and analysis into a benchtop instrument for sample deposition and staining. As shown therein, the microscope and its corresponding optical components are placed around the components of the sample preparation system.

[0077] In some embodiments, the microscope and its optical instruments will be placed below the sample preparation unit of the device (e.g., as shown in FIG. 3). Here, after the sample is deposited on to the glass slide and is stained, the objective unit of the microscope will automatically move closer to the glass slide from below the slide to obtain an image of the relevant portion of the slide. In order to view the glass slide, the movable objective lens will move along the entire length of the slide to obtain the entire image of the slide.

[0078] In some embodiments, the microscope and its optical components are placed above the sample preparation unit of the device. Herein, the objective lens will move closer to the fully prepared glass slide to take the image of the entire glass slide.

[0079] In some embodiments, the imaging modes supported by the microscope shown in FIG. 3 include bright field, dark field, phase contrast and fluorescent microscopy. In other embodiments, the field of view of the glass slide will be projected to the display unit that is present in the system.

[0080] In some embodiments, the microscope shown in FIG. 3 includes features and aspects of the microscope shown in FIGS. 1 and 2.

[0081] In some embodiments, an example workflow for the benchtop instrument shown in FIG. 3 includes the user placing the glass slide and the related consumables into the system, and loading the integrated sample vial into the system. After pressing the start button, the sample will be automatically deposited in the glass slide and stained. After the completion of the sample deposition and staining mode, the instrument will automatically switch to imaging mode. In the imaging mode, the slide will be imaged and projected onto the display that is attached to the instrument. In addition to the on-instrument display, the image will also be projected to a remote computer such that a clinician sitting at a remote site can visually look at the slides in real time.

[0082] One or more of the embodiments described in FIGS. 1 -3 support a variety of different workflows, some of which are illustrated in FIG. 4. As shown therein, a typical workflow that uses the benchtop instrument for deposition and staining with integrated imaging and analysis begins, at operation 410, with dispensing the sample on the slide. In an example workflow, the dispensed sample is then alcohol-fixed and stained using the Papanicolaou stain (or pap stain) in operation 422. In this example, the user (or cytopathologist) has only been tasked with preparing the slide, so the alcohol-fixed and stained slide is sent to the lab in operation 432. As with most medical procedures, results from the laboratory analysis is incorporated into the patient’s Electronic Medical Record (EMR) in operation 440.

[0083] In another example workflow shown in FIG. 4, the sample is dispensed (410), air-dried and stained (424), and immediately reviewed on-site (434) prior to any results being stored in the corresponding EMR (440). Alternatively, the air-dried and stained sample on the slide is subjected to remote evaluation or telepathology (436) prior to EMR storage (440).

[0084] In some embodiments, the operations shown in FIG. 4 can be rearranged based on the clinical requirements. In an example, a sample is dispensed (410), airdried and stained (424), and then send to the lab for analysis (432) prior to being incorporated into the appropriate EMR for that patient (440). In another example, the sample is dispensed (410), alcohol-fixed and PAP stained (422), and then made available for remote evaluation and / or telepathology (436) prior to being incorporated into the appropriate EMR for that patient (440).

[0085] In some embodiments, the remote evaluation and / or telepathology (436) operations can be performed using the independently-configured microscopy system discussed in this patent document. In other embodiments the remote evaluation and / or telepathology (436) operations can be performed using a user interface (e.g., the user interface shown in FIGS. 40A and 40B).

[0086] 2 Examples of benchtop instruments

[0087] FIG. 5 shows an external view of an example benchtop instrument, and FIG. 6 shows a side open-view of the benchtop instrument. As shown in FIG. 6, the benchtop instrument includes the slide processing and clamping modules at the front of the instrument, and an auxiliary systems module at the rear that makes up the main body of the device. The slide processing module receives consumables (hoods and slides) and the SIP consumable from the user and also contains the user interface. The auxiliary systems module contains the electronics, the pneumatics system, and reagent pumps and bottles.

[0088] 2.1 Auxiliary system

[0089] The auxiliary system, shown in FIGS. 7A and 7B, includes the chassis (701 ), the pneumatic system (702), the peristaltic pumps (703), bottles (704), a drip tray (705), the electronics (706), and a power inlet (707).

[0090] 2.1.1 Chassis

[0091] The chassis (701 ) provides the structural framework of the benchtop instrument, and is shown in FIG. 8. As shown therein, the chassis includes the following components:

[0092] - Base plate (801 ), which is machined from a 6mm aluminum plate and provides rigidity to the chassis;

[0093] - Chassis spine (802), which separates the reagent bottles, tubing and pumps from the electronics and the pneumatic system, and provides mounting pointsfor various components of the system. As shown in FIG. 9, the chassis spine separates the fluid side (liquid system and bottles) from the dry side (electronics and pneumatic sub-assembly);

[0094] - Milk crate (803), which is a removable holder for the reagents, buffers and / or fixatives. In an example, the milk crate is designed to hold 250ml bottles for three reagents used in the staining protocol as well as two 15ml vials of buffer and ethanol fixative for the left-hand side;

[0095] - Back plate (804), which is machined from 1 ,6mm steel;

[0096] - False wall (805), which provides a barrier between the reagent bottles and their pumps and tubing. FIGS. 10A and 10B show an example of the false wall (1001 ), which covers the majority of the tubing, thereby imparting a neat appearance to the bottle cavity and allowing it to be easily wiped;

[0097] - Bottle shelf (806), which is staggered such that the tops of the variously- sized bottles are level, and is configured to direct any leakage to the drip tray beneath; and

[0098] - Tube holder (807), which provides hooks to hold the bottle connectors when they are detached. In some embodiments, the position of the tube holder prevents the door from being closed when it is in use, which serves as a reminder to the user to reconnect the bottles.

[0099] 2.1.2 Pneumatic system

[0100] The pneumatic system (702), shown in FIGS. 11 A-11 C, is used in the benchtop instrument for spray deposition. FIG. 11 A shows the configuration and position of an example pneumatic system in benchtop instrument. In some embodiments, the pneumatic system is only pressurized immediately prior to a spray deposition and vented immediately after. In an example, all fittings and tubing are rated to 1 MPa (145.0 psi) and the air pump can achieve a maximum pressure of 0.55 MPa (79.8 psi). Furthermore, all tubing and fittings upstream of the accumulator are sized at 6mm to match the air pump outlets, and all downstream are sixed at 4mm.

[0101] FIG. 11 B shows a rendering of the components of the pneumatic system and FIG. 11C shows the real components used therein. As shown in FIG. 11 B, the pneumatic system includes:

[0102] - Air pump (1 101 ), which is a brushless air pump with a maximum pressure of 550 kPa (79.8 psi) and a flow rate (at atm): 2.7 L / min;

[0103] - Check valve (1102), which prevents air from flowing from the accumulator back towards the pump. In some embodiments, the check valve has tubing with an outer diameter of 6mm and a maximum pressure of 1 MPa (145.0 psi);

[0104] - Accumulator (1103), which is a stainless steel tank that contains a rubber bladder filled with compressed gas. In some embodiments, the accumulator is 6 cubic inches and1 / 8 NPT ports and SMC fittings (e.g., 1x 6mm OD tube, 1x 4mm OD tube). The volume of the accumulator was optimized to provide a consistent pressure during the air pulse while minimizing the footprint;

[0105] - Pressure sensor (1104), which measures from 0 to 1 MPa (145.0 psi) with a 0-5 V analogue output and 4mm OD tube connection interface;

[0106] - Normally open valve (1105), which ensures that the accumulator will vent when the system is unpowered, preventing maintenance on, and transport of, a pressurized system; and

[0107] - Silencer (1106) (or pneumatic muffler), which reduces noise levels and the unwanted discharge of contaminants from the pneumatics.

[0108] 2.1.3 Peristaltic pumps

[0109] The peristaltic pumps (703) deliver the reagents and water for the staining protocol to the right-hand side of the benchtop instrument and remove the waste. In some embodiments, five peristaltic pumps are used to deliver the reagents and water for the Romanowsky staining protocol, and to remove the waste. In some embodiments, the peristaltic pumps are driven by brushed DC motors, controlled by software, and configured to operate in either direction.

[0110] 2.1.4 Bottles

[0111] The bottles are used to house the reagents, fixatives, water, etc. In some embodiments, the bottle capacities are configured based on an expected processing of slides in a day in a cytopathology department.

[0112] In an example, it is assumed that 120 pairs of slides on the benchtop instrument, and the bottle capacities are specified to require the water and waste bottles to be changed once in the middle of the day and the reagents (methanol, ethanol fixative and the stains) and buffer to last an entire day. In this example, the bottle volumes used are:

[0113] - Methanol, Stain 1 and Stain 2: 250ml each;

[0114] - Water, 500ml;

[0115] - Waste, 1000ml;

[0116] - Buffer (e.g., PBS), 15 ml; and

[0117] - Fixative (e.g., CytoFix), 15 ml.

[0118] In some embodiments, the 250ml, 500ml and 1000ml bottles are HDPE for compatibility with the reagents and use no-drip fittings to prevent spillage.

[0119] In some embodiments, the buffer and ethanol fixative are contained in 15ml conical polypropylene tubes.

[0120] 2.1.5 Drip tray

[0121] The drip tray (705) is shown in FIG. 12. As shown therein, the drip tray directs liquids from the bottle shelf (1202) and slide processing overflow port (1203) to the drip tray (1201 ) which can be removed by the user and emptied as required. In an example, the drip tray is fabricated from 1 ,6mm stainless steel.

[0122] 2.1.6 Electronics

[0123] In some embodiments, the electronics (706) run on a 24V DC power supply, and include (i) a first printed broad circuit assembly (PCBA) for the user interface (see § 4.8), including its capacitive touch pads and light-emitting diodes (LEDs) and (ii) a second PCBA containing all the other circuitry required for the software to control the pumps, the valves, the heater, and receive inputs from the system's sensors.

[0124] 2.1.7 Power inlet

[0125] In some embodiments, the benchtop instrument is powered by a 24V DC power supply. In some embodiments, the benchtop instrument may be configured with a hard power switch next to the power inlet (707) on the back, which is connected to a power LED on the user interface at the front of the benchtop instrument.

[0126] 2.2 Slide processing

[0127] FIGS. 13A and 13B show an example of the slide processing mechanism of the benchtop instrument. As shown therein, the slide processing mechanism includes a SIP consumable dock (1301 ), the SIP consumable (1302), a cassette (1303), the hood consumable (1304), slide beds (1305), an ethanol fixative system (1306), a bulkhead (1307), an air spray valve (1308), a buffer system (1309), a slide (1310), an air curtain (1311 ), and a reagent manifold assembly (1312).

[0128] 2.2.1 SIP dock

[0129] The SIP dock, shown in FIGS. 14A-14C (in addition to an orientation detail), locates (or situates) the SIP in the benchtop instrument so that the air nozzles (1402) and spray nozzles (1403) are aligned. Alignment between the air and spray nozzles is important because it affects the strength of the venturi effect on each spray nozzle and, consequently, determines the overall volume of sample sprayed and the evenness of the distribution of the sample (e.g., achieving a monolayer of cells) between the two slides. As shown therein, the SIP dock includes SIP manifold (1401 ), air nozzles (1402), SIP nozzles (1403), SIP side spring (1404), SIP datums (1405), and SIP bed (1406).

[0130] - Air nozzles (1402) are shown in FIG. 15. As seen therein, the air nozzle tips are located 1 ,7mm from the center of the SIP nozzle tips and at an angle of 9.5° to the vertical. While a line drawn to follow this trajectory would land outside the center of the slide (1501 ), testing has shown that the spray 'curves' around the tip of the spray nozzle resulting in a centered spray from this geometry.

[0131] - SIP side spring (1404), shown in FIG. 16A, provides a force (denoted “F” in FIG. 16A) to (i) clamp the SIP nozzles against the SIP body to minimize leakage at points 'A' in FIG. 16A, and (ii) press the left-hand SIP side datum against its hard-stop to accurately locate the SIP nozzle tips under the air nozzles.

[0132] - Operation of SIP switch is shown in FIGS. 16B and 16C. As shown therein, the insertion of the SIP moves the switch rod (1601) to activate the SIP switch (1602). This switch allows the software to determine whether a SIP is present and provide feedback to the user if required.

[0133] - SIP manifold (1401 ), shown in FIG. 17, performs a number of functions.As well as locating the SIP accurately in relation to the air nozzles using the datum features and spring described above, it directs the supply of air from the spray valve (1701 ) to the primary air nozzles (1702) and the air curtain cavities (1703). Plugs are used at either end of the air supply t-section (1704) to enable the air path to be split while still being machined out of a single piece of material.

[0134] 2.2.2 Cassette

[0135] The removable cassette (1303), which is shown in FIGS. 18A-18D, allows access to critical parts of the instrument for cleaning (1801 ). It also contains a number of features to locate the SIP and hood consumables in the instrument.

[0136] In some embodiments, the cassette is replaced with the priming tray (not shown) which catches excess liquids when the priming protocol is run. As shown in FIGS. 18A-18D, the cassette includes the following components:

[0137] - SIP port (1802), which is the hole into which the SIP is inserted is smaller than the thumb guard and, consequently, the SIP cannot be inserted backwards or upside down. The contaminated portions of the SIP (moat and well area) do not come into contact with the cassette or SIP Manifold during insertion or operation. When a SIP is inserted, the curved face surrounding the port combines with the SIP thumb guard to create a torturous path to prevent aerosolized sample from escaping;

[0138] - SIP bed (1803) serves two functions: (i) providing a tactile click to indicate that to the user that they have inserted the SIP sufficiently, and (ii) providing a force on the underside of the SIP nozzles to firmly push its front and top locating features into their reciprocal surfaces on the SIP manifold;

[0139] - Cassette rails (1804), which allow the cassette to be pushed into the instrument by the user and hold it in place during operation;

[0140] - Cassette magnets (1805), which locate and hold the cassette firmly in place in the instrument; and

[0141] - Hood ball detent (1806), which prevents the hood from being removed once the SIP consumable has been inserted into the SIP port.

[0142] 2.2.3 Slide beds

[0143] The benchtop instrument has two slide beds to hold the two slides for 'deposit and stain' processing, or a single slide (on the right) for a 'stain only' protocol. The beds locate the slides accurately under the spray and reagent nozzles, provide an upward spring force to ensure sealing of the hood consumables on the top face of the slides and the right hand bed has a heater to reduce the drying time of that slide.

[0144] In some embodiments, the slides are inserted by hand and easily located.

[0145] After the sample is split and sprayed evenly onto the two slides, the righthand one is dried, fixed, stained and rinsed, while the left-hand one is wet-fixed. The slide beds, shown in FIGS. 19A and 19B, include the following components:

[0146] - Slide beds (1901 ), which are made of machined aluminum, are shown inFIG. 20A. Each slide bed is made of two parts to allow them to be held captive by the housing plate (2001 A).

[0147] - Slide switches (1902) provide an electronic interlock for the air spray valve and reagent pumps to prevent the escape of aerosolized sample and reagent spillages which the absence of slides will cause. The switches also send a signal to the software enabling it to perform checks and provide feedback to the user.

[0148] As shown in FIG. 20B, the slide switches (2002B) are actuated by levers (2003B) which contain spring plungers (2001 B). The plungers increase the range of movement over which the slide switches will be activated allowing a range of slide lengths to be detected. The geometry of the levers reduces the chance of slides being inserted incorrectly, limiting the likelihood of breakage during clamping.

[0149] - Slides (1903) meet the ISO 8037 / 1 -1986 specification:Length:Width: Thickness:

[0150] Herein, the subscript and superscript represent the specified tolerance.

[0151] - Overflow drain (1904) guides fluids to the spill tray underneath the reagent bottles in the event of a leak in the slide processing area.

[0152] - Slide bed springs (1905) are shown in FIG. 21 . As shown therein, two springs (2102, 2103) under each slide bed (2101) fulfill two functions: (i) providing an even clamping force between the hood consumable seal and the slide face to prevent the escape of aerosolized sample and reagents, and (ii) allowing the instrument to accommodate a range of slide thicknesses.

[0153] In some embodiments, the left-hand slide bed has a lighter spring force as it only needs to seal against aerosolized sample. This is provided by two spring plungers located centrally, front and back. The right-hand slide bed requires a stronger force to seal against the reagents pumped onto the slide during staining and has two ball detents located diagonally opposite each other. Their position is restricted by the heater matt, also located on the underside of the slide bed.

[0154] - Heater (1906) is adjacent to the right-hand slide bed, which is heated to dry the slide before fixing and staining. The heater is a custom heater mat assembly which contains: (i) 20W (@24VDC) heater mat, (ii) twin thermistors to provide feedback to the temperature control system, and (iii) potted wiring. The assembly is fixed to the bottom part of the slide bed with epoxy.

[0155] 2.2.4 Ethanol fixative system

[0156] In some embodiments, and as shown in FIG. 22, fixative spray system is comprised of a 15ml reservoir of an ethanol-based fixative (e.g., CytoFix), a high- frequency dosing pump (2201) and a custom brass insert with aerosolizing features (2203) which is attached to the SIP Manifold and supplied by paths within it (shown on right-hand side figures). The ethanol fixative system includes an ethanol pump (2201 ), an inlet fitting (2202, and shown in FIG. 23), and an ethanol spray insert (2203).

[0157] - Ethanol spray insert (2203) contains a brass fitting containing features to aerosolize the small (~400pL) volume of ethanol supplied through the SIP manifold. In some embodiments, it is located against the ethanol supply hole on the SIP manifold by the countersunk screw holes.

[0158] As shown in FIGS. 24A and 24B, ethanol flows in at (2401 ), passes along the 0.5mm wide passage to the swirl feature (2402) and out through the 0.3mm nozzle. The cutout feature (2404) allows the nozzle length to be kept short (0.24mm) with clearance around the exit to allow the spray to fan outwards.

[0159] 2.2.5 Bulkhead

[0160] In some embodiments, the bulkhead (1307) is machined from a 6mm aluminum plate for strength and rigidity. In an example, it is fixed by its mounting brackets to the baseplate at a 4° angle to the vertical to facilitate draining of reagents and water to the scavenger port on the right-hand side.

[0161] 2.2.6 Air spray valve

[0162] In some embodiments, the air spray valve (1308) is closed until it receives 24V DC signal from the software-controlled electronics. The benchtop instrument is designed to accommodate spray bursts of 30 to 100ms in length.

[0163] 2.2.7 Buffer system

[0164] In some embodiments, and as shown in FIG. 25, the buffer system is comprised of a 15ml reservoir of PBS buffer, a micro-dosing pump (2501 ) and the buffer nozzle (2502). Holes in the SIP manifold supply the pump with buffer and move the liquid from its outlet to the nozzle. The pump-to-nozzle fluid path uses a plug to enable machining out of a single piece. The buffer system includes:

[0165] - Buffer pump (2501 ), which deposits, via the buffer nozzle, 10pL of sample into the SIP well prior to the spraying of sample onto the slides. In an example, buffer pump is calibrated to 10pL during manufacture.

[0166] - Buffer nozzle (2502) is located above the sample will in FIG. 26.

[0167] 2.2.8 Air curtain

[0168] In some embodiments, and as shown in FIGS. 27A and 27B, the brass inserts on the outer sides of the air nozzles (2701 ) create a 'curtain' of air (1311 ) which prevents the sample spray (2704) landing on the hood consumable collars, increasing the yield of sample on the slides. In an example, each insert contains 5x 0.6mm diameter holes 2.2mm long (2702). The diameter of these was set to achieve a nominal SIP vacuum of 7kPa with a supply pressure of 35psi. The air curtain holes are fed from a chamber on the sealing side of the insert (2703) and coincident face on the SIP manifold.

[0169] 2.2.9 Reagent manifold assembly

[0170] In some embodiments, and as shown in FIGS. 28A and 28B, the reagent manifold assembly (1312) contains the reagent manifold itself - which supplies reagents to, and drains them from, the right-hand slide - as well the priming tray switch, hood switch rods and switches and the SIP switch.

[0171] - Priming tray switch (2801 ), which informs the software whether the priming tray is in place to avoid the priming protocol running in its absence.

[0172] - Hood switches (2802), shown in FIGS. 29A and 29B, include one rod per hood (2901 ), which engages the switches when the hood is inserted into the instrument, a hood switch actuator (2902), which depresses the pair of microswitches for a given hood, and two microswitches for redundancy (2903), which provide an electronic interlock for the air spray valve and reagent pumps to prevent the escape of aerosolized sample and reagent spillages which the absence of hoods will cause. The switches also send a signal to the software enabling it to perform checks and provide feedback to the user.

[0173] - Scavenger port (2803), shown in FIGS. 30A-30C, drains reagents and water from the right-hand slide. The bulkhead, and consequently the slides, are on a 4° angle to the horizontal which means that fluids run towards the scavenger port.

[0174] The hood consumable (3001 ) is inserted when the instrument is in the unclamped state and comes to a stop (assisted by the detent feature) with its locating feature (3002) against the hard stop face of the reagent manifold (3003).

[0175] When the instrument is clamped by the user the scavenger feature (3004) and reagent nozzles (3007) fit into the hole on the hood. The circular face of thescavenger port (3005) is located 0.5mm above the surface of the slide. As this gap is emptied of liquid by the drain peristaltic pump via the drain hole (3006), adjacent liquid flows in, due to the capillary effect, to take its place. This flat surface is circular so that it drains evenly from all sides, maximizing the liquid removed from the slide.

[0176] - Reagent paths (2804) are shown in FIGS. 31 A and 31 B. Reagents enter the manifold via tubing from the peristaltic pumps. The methanol and red stain paths within the manifold are shared due to limited space and their being the most compatible combination of the 4 liquids. Methanol and the two stains exit the manifold via nozzles, whereas the water exit port is designed to run water down the scavenger feature during dispensation to remove any splashes of stain and increase the final cleanliness of the slide.

[0177] 2.3 Clamping

[0178] FIG. 32A shows the components of an example clamping mechanism used in the benchtop instrument, which includes a handle (3201 ), a slide bed (3202), a solenoid latch (3203), a clamping interlock with two microswitches (3204), and a damper (3205). The operation of the clamping mechanism is described first, and then the particulars of certain components are discussed.

[0179] In some embodiments, the handle (3201) connected at the slide bed (3202) is pushed downwards by the user moving, via the solenoid latch (3203) linkages, the slide bed (3202) upwards along its linear rail. As the slides rise they push the hoods up against the underside of the cassette which, as they reach their clamping surfaces in turn compress the slide bed springs and seal against the slides. The solenoid latch (3203) fasten locks the mechanism in place until the solenoid is activated by the software-controlled electronics at which time the slide bed lowers gently - slowed by the rotary damper engaging with the vertical rack teeth.

[0180] The two microswitches of the clamping interlock (3204) are activated when the mechanism is in the clamped position. These provide an electronic interlock for the air spray valve and reagent pumps to prevent the escape of aerosolized sample and reagent spillages which running the instrument unclamped will cause. The switches also send a signal to the software enabling it to perform checks and provide feedback to the user. Two switches are used for redundancy.

[0181] 2.3.1 Handle

[0182] An example of the handle (3201 ) is shown in FIG. 32B on the face of the benchtop instrument. It may be made from anodized aluminum and can be configured to operate vertically on a linear rail.

[0183] 2.3.2 Solenoid latch

[0184] In some embodiments, the solenoid latch (3203), which is detailed in FIG. 33, holds the clamping assembly closed until the solenoid (3302) is activated by the software. The height of the latch (3301 ) is calibrated during the manufacturing process to fix the clamping force of the hood consumables against the slides.

[0185] 2.3.3 Damper

[0186] In some embodiments, the damper (3205) slows the opening, under gravity, of the clamping mechanism when the latch releases it.

[0187] 2.3.4 Clamping interlock

[0188] In some embodiments, the clamping interlock (3204) includes two microswitches (for redundancy) to ensure spray and stain valve and pumps will not operate unless the clamping mechanism is engaged to prevent spillage of reagents and escape of aerosolized sample.

[0189] 2.4 Covers

[0190] FIG. 34 shows examples of covers for the benchtop instrument. As shown therein, the portions that cover the inner components and circuitry include a side-panel (3401 ) and a door (3402) on one side of the benchtop instrument, another side-panel (3403) on the other side of the benchtop instrument, and a front cover (3404).

[0191] In some embodiments, the front cover is made from vacuum-cast plastic, and the side covers and doors are fabricated using sheet metal (e.g., 1 ,6mm steel stock). In other embodiments, the different combinations of plastic and sheet metal may be used to design the covers and doors for the benchtop instrument.

[0192] 2.5 Door

[0193] The benchtop instrument includes a door that enables operators and / or technicians to access the internal components and circuitry of the instrument. In some embodiments, and as shown in FIG. 35A, friction hinges (3501 ) can be used to ensure that the door can be opened to any position with no risk of falling closed.

[0194] In some embodiments, and as shown in FIG. 35B, dual microswitches (3502) are activated when the door is in the closed position. These provide an electronicinterlock for the reagent pumps to ensure that the reagent tubing can only fail under pressure with the door closed, protecting the user and operating environment from reagents. The switches also send a signal to the software enabling it to perform checks and provide feedback to the user. Two switches are used for redundancy.

[0195] 2.6 Handles

[0196] The benchtop instrument includes a front handle and a rear handle. In some embodiments, and as shown in FIGS. 36A and 36B, the rear handle (3601) is set high in the backplate to maximize stability when lifting the benchtop instrument and the front handle (3602) is incorporated into the front fascia. In an example, the rear handle may be an original equipment manufacturer (OEM) part and the front handle may be a custom-designed part. In other examples, both the front and rear handles may be OEM parts or both may be custom designed.

[0197] 2.7 Feet

[0198] In some embodiments, the benchtop instrument has four non-slip feet that are chemically compatible with the reagents being used.

[0199] 2.8 User interface

[0200] In some embodiments, the user interface is incorporated into the front fascia, and includes four capacitive touch buttons and a number of status LEDs.

[0201] 3 Examples of ML-based imaging and analysis

[0202] In some embodiments, the benchtop instrument includes a microscopy subsystem that is configured to rapidly provide a digitized representation of the slide with the deposited and stained sample for both on-site and remote analysis by the (cyto)pathologist and / or the medical team. In other embodiments, the imaging subsystem is functionally integrated with (but physically separated from) the benchtop instrument, and also supports both on-site and remote analysis.

[0203] In some embodiments, the microscopy subsystem includes a remotely- operable microscope and a digital slide (pathology) scanner. The remotely-operable microscope has a completely-automated X, Y, and Z actuation stages to automatically control the movement of the specimen slide. It further includes automated turrets and a camera to control both the objective lenses and visualize the images in a computer. The remotely-operable microscope is typically kept in places where patient procedures are performed (e.g., during Mohs surgery, Rapid On-site Evaluation for bronchoscopy, etc.), and can be controlled using screen-sharing software (e.g., TeamViewer, MicrosoftTeams, etc.). In this example, a pathologist (or cytopathologist) can connect to the remotely-operable microscope using one of the screen-sharing software, examine the specimen slides remotely, and provide a diagnostic evaluation to the physicians.

[0204] In some embodiments, the microscopy subsystem includes a digital scanning element that automatically scans the entire slide at a low magnification (e.g., 10 X). The microscope supports higher magnification objectives (e.g., 40 X or above) that are primarily used for remote viewing of the specimen slide. The microscopy subsystem can be operated in dual modes: (1 ) the slide scanning mode and (2) the (remote) viewing mode.

[0205] In some embodiments, the microscopy subsystem includes completely- automated X, Y, and Z actuation stages, as well as automated turrets, which enable the microscope to be quickly toggled between slide scanning and viewing modes.

[0206] In the described embodiments, one or more machine-learning (ML)-based imaging and analysis algorithms are used to analyze the digitized images generated by the microscopy subsystem. In some examples, as shown in FIG. 37, an ML algorithm module 3720 can include a model training engine, a model evaluation engine, and a result processing and distribution engine. As shown in FIG. 37, training / testing data that is generated (e.g., by performing one or more pre-processing algorithms described later) is stored in the training / testing dataset 3710. The machine learning algorithm module 3720 includes the model training engine (e.g., which is trained using the training / testing dataset 3710), the model evaluation engine, and the result processing engine, which processes and formats the output such that it can be utilized and displayed in a user interface 3730.

[0207] In some embodiments, an example workflow for the microscopy subsystem and the ML-based imaging and analysis algorithms includes:

[0208] - A specimen slide is placed on the microscope, and its presence is automatically sensed (e.g., using a weight sensor or an optical sensor), which triggers obtaining a macroscopic view of the entire slide image (e.g., using the 10X objective lens). Examples of the macroscopic view for two deposited and stained slides 3810 and 3910 are shown in FIGS. 38A and 39A, respectively.

[0209] - An ML-based Specimen Region Identification (SRI) algorithm is executed, which identifies the region within the specimen slide where the specimen is deposited. Examples of the region (3820 and 3920) within the specimen slide (3810 and3910) being identified are shown in FIGS. 38B and 39B, respectively. The coordinates on the slide where the specimen is located are automatically identified, and the microscopy subsystem switches to the slide scanning mode. In some embodiments, the SRI algorithm operates with a predefined probability of detection when identifying the region within the specimen slide that contains the deposited specimen.

[0210] - The slide scanning mode automatically activates the X, Y, and Z actuation stages to start scanning the region of the specimen slide corresponding to the coordinates that were received from the SRI algorithm.

[0211] - As the specimen slide is getting scanned, the ML-based RelevantTile Identification (RTI) algorithm analyzes the scanned region at a finer degree of granularity (referred to as a “tile”).

[0212] In some embodiments, the RTI algorithm identifies whether a tile of interest contains any relevant cells (e.g., epithelial cells, lymphocytes, etc.). The RTI algorithm automatically flags any identified tiles and saves the location of these tiles in order for the pathologist or cytopathologist to examine those tiles in real time. Examples of identified tiles (3820-1 , 3820-2, 3820-3, ..., 3820-N and 3920-1 , 3920-2, 3920-3, ..., 3920-N) within the specimen slide (3810 and 710) being identified are shown in FIGS. 38C and 39C, respectively.

[0213] In some embodiments, the tile size used in the RTI algorithm is based on the magnification of the lens used and the field of the camera. In an example, if the tile size is too small (e.g., due to a small field of view camera (or sensor size of the camera) or a large magnification objective lens), then the microscope will take a longer time to scan the entire specimen slide due to the number of steps it takes to scan a 25 mm x 50 mm area. In another example, if the tile size is too big (e.g., due to a very large field of view camera or a very small magnification objective lens), then the objects within the tile will be too small to be discerned by the RTI algorithm. The described embodiments typically use a tile size between 300pm and 1500pm, with the specific size being selected based on, for example, the size of the cells of interest.

[0214] In some embodiments, the RTI algorithm identifies tiles of interest that contain relevant cells with a predefined probability of detection (e.g., determined and calibrated using training data). In some examples, the probabilities of detection of the SRI and RTI algorithms are selected independently. In other examples, the RTI’s probability of detection is greater than or equal to the probability of detection of the SRIalgorithm. In yet other examples, the SRI’s probability of detection is greater than or equal to the probability of detection of the RTI algorithm.

[0215] - The user interface of the system (e.g., user interface 3730 in FIG.37) is configured to display relevant tiles to a pathologist in real-time.

[0216] - The pathologist can switch from the slide scanning mode to the viewing mode using a one-click selection control on the user interface. In the viewing mode, the pathologist can switch to a higher objective lens (e.g., 40X or higher) to observe the relevant cells of interest at a higher magnification and a higher resolution.

[0217] In contrast with existing systems that typically require around 5-10 minutes to scan the entire specimen slide, the described embodiments can extract relevant information from the specimen slide (e.g., are there any cells of interest?) using the dual scanning and viewing modes, and the described ML-based SRI and RTI algorithms, in times that are on the order or 30 seconds or less (and that are readily available to both on-site and remote pathologists). Furthermore, the need to move the microscope using a mouse- or keyboard-controlled interface is obviated, as is a need for performing a manual scan of any region of interest on the specimen slide.

[0218] As previously discussed, embodiments of the disclosed technology provide the ability to configure the microscopy hardware and ML-based software system as an independent tool for cytological and cytopathological analysis (e.g., as shown in FIGS. 1 and 2, physically separated from the benchtop instrument), as well as a microscopy (sub)system (which includes a microscope element and a slide scanning element) that is integrated with the benchtop instrument (e.g., as shown in FIG. 3). In some examples, the microscopy (sub)system (which includes one or more processors, a slide scanner, and a microscope) is configured to implement a method for analyzing a cellular sample on a substrate (e.g., a specimen slide) using, for example, the SRI and RTI algorithms as described above.

[0219] As is described in Section 2, the design of the spray nozzle (and other factors) affects the footprint of the specimen deposited on the slide (or substrate). Thus, the SRI algorithm is also configured to receive as inputs, one or more parameters related to the hardware specification and or design, which enables the SRI algorithm to locate the region of interest more rapidly. In some examples, the one or more parameters include distance of spray nozzle from substrate, air pressure, air flow rate, spray nozzle design, per-spray volume, total volume, drying temperature, air dryingtime, cell type, cell concentration, and / or transport or suspension liquid characteristics. These (non-limiting) factors can be selected to configure the footprint of the spray deposition on the slide / substrate, and can consequently be used by the SRI algorithm, e.g., to determine a starting coordinate (or ending coordinate) for the scan.

[0220] In some embodiments, the RTI algorithm analyzes tiles in the scanned region based on a predefined scanning configuration. In some examples, the scanning configuration specifies the size of the tile, the magnification of the objective lens used, the order in which the tiles are scanned (e.g., horizontal sweep, vertical sweep, snake scan, i.e. , alternating directions, inward spiral, outward spiral, etc.), and parameters for the identification and classification algorithms.

[0221] Embodiments of the disclosed technology provide an ML-based system that includes, in an example, a data preparation process (or module) and an ML algorithm module (e.g., as shown in FIG. 37).

[0222] In some embodiments, the data preparation module is configured to implement one or more of the following pre-processing methods for the input data (e.g., used in training dataset 3710 in FIG. 37).

[0223] - Thresholding method. Can be size-based or intensity-based. A sizebased thresholding algorithm is used to remove debris and other tiny structures from the slide images, and an intensity-based thresholding algorithm is used to remove artifacts from the slide images.

[0224] - Opening and closing methods. These methods are used to identify clusters of cells within the slide images. In general, the cells of interest are present in groups or clusters while the red blood cells are present throughout the slide as individual cells. Thus, the opening and closing techniques are used to pre-process the slides to extract these groups of cells. In an example, the opening and closing methods use one or more of K-means clustering, density-based spatial clustering, a Gaussian mixture model based clustering, a Balance Iterative Reducing and Clustering using Hierarchies (BIRCH) clustering algorithm, an affinity Propagation clustering algorithm, a mean-shift clustering algorithm, or an agglomerative hierarchy clustering algorithm.

[0225] - Surface roughness methods. The surface roughness, which can be calculated as the variance of intensities within a group of pixels, is used to identify the types of cells. In general, red blood cells have a smoother cell surface which in turn hasa lower surface roughness value compared to epithelial cells. Thus, this metric can be used to identify the cells of interest from a red blood cells background.

[0226] - Random flip method. Randomly flip the orientation of an input image(horizontally and vertically). This is primarily done to force the model to look at the entire input image instead of looking into the same spot in an image every time during the training process.

[0227] - Random rotation method. Randomly rotate the input image to prevent the training algorithm from looking at the same spot in an image.

[0228] - Rescaling method. Rescale input image values to [0,1 ] or [-1 ,1] depending on transfer learning model’s desired input. This is done to rescale the image regardless of the types of imaging sources that are used to obtain the input image.

[0229] - Image sharpening method. This is done by passing a kernel over the image to “sharpen” the borders of cells / objects in the input image. This methodology is primarily used to overcome the fuzziness in an input image due to problems encountered with the imaging system.

[0230] - Resizing method. A single model or multiple parallel models are created to account for training images of different image sizes (256x256x3, 512x512x3, etc.). This is done to extract different details from the training image. The higher resolution model with a small image size will extract finer details like the size and shape of small cells (e.g., lymphocytes) while the lower resolution model with a large size will generalize details more.

[0231] - Noise model. Add random noise to force the training model to not learn too quickly from inputs that have similar patterns but are not generalizable to the rest of the dataset. The purpose of this method is similar to the random flip method and the random rotation method to make sure that the model is forced to look at different features within the input image.

[0232] - Contrast normalization method. The global and local contrast normalization is performed to reduce the overall contrast in input image. While this method is counter-intuitive, the purpose of reducing the contrast of the entire training image or a subsection of the training image is so that smaller objects can be better visualized. For example, if there is a very small contrast between the small cells and its background, reducing the overall contrast of the image will help to better visualize the cells in real-time usage of the machine.

[0233] - Color channel method. Most images in the dataset contain only red or blue colors (or both). This is due to the type of staining that is used to stain the cells in the images. Hence, removing the green layer could force the machine to focus on red and blue colors more while also reducing input size for faster classification.

[0234] In some embodiments, the machine learning algorithm module 3720 in FIG. 37 is configured to automatically identify diagnostic cells from the background of artifacts and red blood cells. In an example, the module includes the following:

[0235] - 21 ,802,784 total parameters,

[0236] - 94 total convolution layers,

[0237] - 94 total batch normalization layers,

[0238] - 94 total activation layers,

[0239] - 15 concatenate layers,

[0240] - 9 two-dimensional average pooling layers,

[0241] - 4 two-dimensional max pooling layers, and

[0242] - 1 input layer.

[0243] In addition to the above base model, some model configurations include:

[0244] - a two-dimensional global average pooling layer,

[0245] - converting base model output to a column vector for the next layer,

[0246] - one dense layer with four nodes, with each node corresponding to one of the four classes discussed below, and

[0247] - returning a four-dimensional column vector (1 dimension for each class), and containing, for example, 8196 parameters.

[0248] In some embodiments, the above-described module is trained to classify the following four classes: (1 ) red blood cells, (2) background glass slide, (3) artifacts, and (4) diagnostic epithelial cells. In some embodiments, the algorithms are configured to maximize the identification of the diagnostic epithelial cells. In other embodiments, the above-described module is trained to classify between the following four classes: (1 ) benign cells, (2) malignant cells, (3) atypical cells, and (4) infectious cells. In the described embodiments, the training model is still utilized to identify the 4 classes (in either / both cases) so that a new training of all the input images is not required for further classification of cells in the current data.

[0249] In some embodiments, the Adam optimization algorithm is used to iteratively update network weights in training data. In other embodiments, a stochastic gradient descent (SGD) method is used.

[0250] In some embodiments, the result processing engine of the machine learning algorithm module 3720 in FIG. 37 is configured to implement one or more of the following post-processing methods on the output of the model.

[0251] - Application of SoftMax function. The SoftMax function (or normalized exponential function) is applied to the model output to generate the probability distribution of the different classes discussed above. That is, the SoftMax function is the last activation function that normalizes the output of the network to a probability distribution over predicted output classes.

[0252] - Classification output. In an example, once the probability distribution has been generated using the SoftMax function, the index of the maximum value based on the results of the probability distribution function is determined. In the abovedescribed example, a given image will be classified as either a glass slide, an artifact, red blood cells, or diagnostic epithelial cells based on the results of the probability density function. Alternatively, only the images that either have red blood cells or diagnostic epithelial cells can be displayed to the end user, without displaying an image with artifacts or the glass slide.

[0253] - Thresholding classified output. In some embodiments, only those cells in which the number of diagnostic epithelial cells or red blood cells are above a certain minimum number are displayed to the end user. For example, if only 1 -2 diagnostic epithelial cells are identified, then the corresponding image will not be displayed. This advantageously ensures that there are always enough cells to make a robust and informed decision with a relatively high degree of confidence.

[0254] Embodiments of the disclosed technology further provide a user interface (III), as shown in the examples in FIGS. 40A and 40B, which correspond to screenshots taken at a first time and a second time, respectively. As shown therein, the user interface supports the following features.

[0255] [A] Display for the macroscopic view of the specimen slide obtained using the microscope (e.g., as shown in FIGS. 38A and 39A).

[0256] [B] Controls for the X, Y, and Z actuation stages that enable the user to manipulate the slide scanner.

[0257] [C] Controls for the microscope, which include “Auto Focus” and different magnifications, e.g.4X or 40X.

[0258] [D] Display for slide scanner progress. The size of the overall region is indicative of the result of the SRI algorithm, which identifies the portion of the slide that includes samples of interest. This portion of the Ul also shows the progress of the RTI algorithm, which is scanning tiles while the SRI algorithm is still executing, i.e. , in a pipelined manner. Additionally, the dark bolded outline for a tile is indicative of that being a tile that includes cells of interest. Since the screenshot in FIG. 40B was taken after the screenshot in FIG. 40A, the number of flagged tiles happens to have in increased in this example.

[0259] [E] Display for parts of one or more flagged tiles that have been explicitly tagged by the user after visual analysis at a higher magnification.

[0260] [F] The higher magnification view of the portion of a flagged tile.

[0261] In some embodiments, the user interface shown in FIGS. 40A and 40B can be used to identify cell types using the higher magnification view of the portion of the flagged tile. In some examples, the identified cell type includes a benign cell type, a malignant cell type, an atypical cell type, and / or an infectious cell type. In other examples, the identified cell type includes red blood cells or diagnostic epithelial cells. In yet other embodiments, the identified cell type includes a benign cell type, a malignant cell type, an atypical cell type, an infectious cell type, a red blood cell type, and / or a diagnostic epithelial cell type. In these examples, the cell type is differentiated (or identified) from the background of the slide and / or other artifacts present on the slide.

[0262] 4 Example embodiments of the disclosed technology

[0263] The described embodiments provide a mobile, dedicated system to perform rapid on-site evaluation of biological samples. Advantages of the benchtop instrument include, inter alia, fine needle aspiration (FNA) biopsy adequacy, the ability to triage the specimen, shorten the length of the procedure, help guide and direct the biopsy in real time, and overall assist in managing the patient’s procedure with the performing clinician at the point of care. In particular, the problem being solved by the disclosed technology is the current inability to consistently produce a monolayer of cells on a substrate, which can then be analyzed by a cytopathologist in a timely manner, at the point of care.

[0264] FIG. 41 shows a flowchart for an example method of analyzing a cellular sample on a substrate. As shown therein, the method 4100 includes, at operation 4110,generating, at a first magnification level, a macroscopic view of a surface of the substrate. In some examples, the macroscopic view is generated using a microscope physically separated from the benchtop instrument (e.g., shown in FIGS. 1 and 2) that was used to produce the substrate. In other examples, the microscope both functionally and physically integrated with the benchtop instrument (e.g., shown in FIG. 3).

[0265] The method 4100 includes, at operation 4120, identifying, using one or more processors, coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest. Herein, the one or more processors are implementing the Specimen Region Identification (SRI) algorithm with a first probability of detection.

[0266] The method 4100 includes, at operation 4130, generating, at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view. In some embodiments, the first digitized representation is generated using a slide scanner, which is configured to work in conjunction with the microscope.

[0267] The method 4100 includes, at operation 4140, identifying, using the one or more processors, one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection. Herein, the one or more processors are implementing the Relevant Tile Identification (RTI) algorithm with a second probability of detection.

[0268] The method 4100 includes, at operation 4150, generating, using the slide scanner at a third magnification level, a second digitized representation of at least one tile of the one or more tiles. In some embodiments, the second digitized representation is generated using the slide scanner. In other embodiments, the second digitized representation is generated in response to a user request comprising an identifier of the at least one tile. In some embodiments, the first and second magnification levels are identical, and the third magnification level is either the same or greater than the first and second magnification levels. In other embodiments, the first magnification level, the second magnification level, and the third magnification level are all different.

[0269] In some embodiments, method 4100 further includes receiving another user request comprising a value for the third magnification level, and the slide scanner is configured to switch from the second magnification level (which may be identical to the first magnification level) to the different (and higher) third magnification level.

[0270] The method 4100 includes, at operation 4160, identifying, based on the second digitized representation, at least one cell type within the samples of interest. In some embodiments, the at least one cell type that is identified includes a benign cell type, a malignant cell type, an atypical cell type, and / or an infectious cell type. In other embodiments, the at least one cell type can further include a red blood cell type and / or a diagnostic epithelial cell type. In these embodiments, the at least one cell type is differentiated (or identified) from the background of the specimen slide and / or other artifacts on the specimen slide.

[0271] Embodiments of the disclosed technology provide, in some aspects, the following technical solutions:

[0272] Solution 1 . An apparatus for depositing, staining, and analyzing a cellular sample, comprising: a slide processing module, positioned at an upper front position of the apparatus, comprising: a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample, and a clamping module positioned at a lower front of the apparatus; an auxiliary systems module, positioned at a rear of the apparatus, comprising: a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol-based fixative bottle, and an electronics subsystem configured to execute, based on the input from the user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on a surface of a substrate; and an optical imaging and analysis system, configured to analyze the stained cellular sample, comprising: a microscopy subsystem configured to generate a digitized representation of the surface of the substrate at one or more magnification levels, and one or more processors configured to identify one or more portions of the digitized representation that include the stained cellular sample.

[0273] Solution 2. The apparatus of solution 1 , wherein the microscopy subsystem comprises a microscope that operates independently of the slide processing module.

[0274] Solution 3. The apparatus of solution 2, wherein the microscope is located adjacent to and separate from the apparatus.

[0275] Solution 4. The apparatus of solution 2, wherein the optical imaging and analysis system supports telepathology.

[0276] Solution 5. The apparatus of solution 1 , wherein the optical imaging and analysis system is positioned below the slide processing module.

[0277] Solution 6. The apparatus of solution 1 , wherein the optical imaging and analysis system is positioned above the slide processing module.

[0278] Solution 7. The apparatus of any of solutions 1 to 6, wherein the optical imaging and analysis system is configured to support bright field, dark field, phase contrast, and fluorescent microscopy imaging modes.

[0279] Solution 8. The apparatus of any of solutions 1 to 7, wherein the one or more processors is configured to: perform at least one pre-processing technique on the digitized representation.

[0280] Solution 9. The apparatus of solution 8, wherein the at least one preprocessing technique comprises a thresholding technique, an opening and closing technique, a surface roughness technique, a random orientation technique, an image sharpening technique, or a contrast normalization technique.

[0281] Solution 10. The apparatus of solution 8 or 9, wherein the one or more processors is further configured to perform, subsequent to the at least one preprocessing technique, a machine-learning (ML) classification algorithm on the digitized representation.

[0282] Solution 11 . The apparatus of solution 10, wherein performing the ML classification algorithm comprises identifying the one or more portions of the digitized representation.

[0283] Solution 12. The apparatus of solution 10 or 11 , wherein the ML classification algorithm is configured to detect diagnostic epithelial cells amidst red blood cells, artifacts, and glass slides.

[0284] Solution 13. The apparatus of any of solutions 1 to 12, wherein the microscopy subsystem comprises a microscope and a slide scanner, each operable at the one or more magnification levels.

[0285] Solution 14. The apparatus of solution 13, wherein the optical imaging and analysis system is configured, as part of analyzing the stained cellular sample, to: generate, using the microscope at a first magnification level, a macroscopic view of the surface of the substrate; identify, using the one or more processors, coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection; generate, using the slide scanner at the first magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view; identify, using the one or more processors, one ormore tiles in the first digitized representation that include the samples of interest with a second probability of detection; and generate, using the slide scanner at a second magnification level higher than the first magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile.

[0286] Solution 15. The apparatus of solution 14, wherein the one or more tiles being identified begins prior to the first digitized representation being generated completes.

[0287] Solution 16. The apparatus of solution 14, wherein the user request is from a remote user.

[0288] Solution 17. The apparatus of solution 14, wherein an indication of whether the substrate comprising the stained cellular sample warrants further processing is based on a number of the one or more tiles.

[0289] Solution 18. A method of analyzing a cellular sample on a substrate, comprising: generating, using a microscope at a first magnification level, a macroscopic view of a surface of the substrate; identifying, using one or more processors, coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection; generating, using a slide scanner at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view; identifying, using the one or more processors, one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection; generating, using the slide scanner at a third magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile; and identifying, based on the second digitized representation, at least one cell type within the samples of interest.

[0290] Solution 19. The method of solution 18, wherein the one or more processors, the microscope, and the slide scanner are components of a cytopathology imaging and analysis unit.

[0291] Solution 20. The method of solution 18 or 19, wherein the first probability of detection is configured based on a first set of training data for a specimen region identification algorithm, and wherein the second probability of detection is configured based on a second set of training data for a relevant tile identification algorithm.

[0292] Solution 21 . The method of solution 20, wherein the first probability of detection is greater than or equal to the second probability of detection.

[0293] Solution 22. The method of solution 20, wherein the first probability of detection is less than or equal to the second probability of detection.

[0294] Solution 23. The method of any of solutions 18 to 22, wherein the cellular sample on the substrate is produced using a benchtop instrument for deposition and staining.

[0295] Solution 24. The method of solution 23, wherein the benchtop instrument comprises: a slide processing module, positioned at an upper front position of the benchtop instrument, comprising: a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample, and a clamping module positioned at a lower front of the benchtop instrument; and an auxiliary systems module, positioned at a rear of the benchtop instrument, comprising: a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol-based fixative bottle, and an electronics subsystem configured to execute, based on the input from the user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on the surface of the substrate.

[0296] Solution 25. The method of solution 24, wherein the cellular sample is in a transport solution, and wherein the deposition operation comprises spraying the cellular sample in the transport solution from a spray nozzle onto the surface of the substrate.

[0297] Solution 26. The method of solution 25, wherein identifying the coordinates of the portion of the macroscopic view or identifying the one or more tiles in the first digitized representation is based on at least one of a design of the spray nozzle, a distance between the spray nozzle and the surface of the substrate, a cell type or a cell concentration of the cellular sample, or a characteristic of the transport solution.

[0298] Solution 27. The method of any of solutions 18 to 26, wherein the first magnification level is identical to the second magnification level.

[0299] Solution 28. The method of solution 18 or 27, wherein the third magnification level is greater than the first magnification level and the second magnification level.

[0300] Solution 29. The method of solution 18 or 27, wherein the third magnification level is identical to the second magnification level.

[0301] Solution 30. The method of any of solutions 18 to 29, wherein the at least one cell type comprises a benign cell type, a malignant cell type, an atypical cell type, or an infectious cell type.

[0302] Solution 31 . The method of any of solutions 18 to 29, wherein the at least one cell type comprises a red blood cell type or a diagnostic epithelial cell type.

[0303] Solution 32. A system for analyzing a cellular sample on a substrate, comprising: a microscope configured to generate a macroscopic view of a surface of the substrate at a first magnification level of a set of magnification levels; a slide scanner, coupled to the microscope, configured to generate a digitized view of the surface of the substrate at each of the set of magnification levels; and one or more processors, coupled to the microscope and the slide scanner, configured to identify coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection; wherein the slide scanner is further configured to generate, at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view, wherein the one or more processors is further configured to identify one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection, wherein the slide scanner is further configured to generate, at a third magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile, and wherein the one or more processors is further configured to identify, based on the second digitized representation, at least one cell type within the samples of interest.

[0304] Solution 33. The system of solution 32, wherein the first magnification level is identical to the second magnification level and the third magnification level.

[0305] Solution 34. The system of solution 32, wherein the third magnification level is greater than the first magnification level and the second magnification level.

[0306] Solution 35. The system of solution 32, wherein the at least one cell type comprises a benign cell type, a malignant cell type, an atypical cell type, an infectious cell type, a red blood cell type, or a diagnostic epithelial cell type.

[0307] Solution 36. The system of any of solutions 32 to 35, further comprising: a benchtop instrument, comprising: a slide processing module, positioned at an upper front position of the benchtop instrument, comprising: a user interface configured to receive an input from a user that configures a deposition operation and a stainingoperation for the cellular sample, and a clamping module positioned at a lower front of the benchtop instrument; and an auxiliary systems module, positioned at a rear of the benchtop instrument, comprising: a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol- based fixative bottle, and an electronics subsystem configured to execute, based on the input from the user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on the surface of the substrate.

[0308] Solution 37. The system of any of solutions 32 to 36, comprising: a user interface comprising: a first portion configured to display a digitized representation of the macroscopic view of the surface of the substrate; a second portion configured to display the second digitized representation of the at least one tile; a first set of controls configured to control a movement of the slide scanner in an x-axis direction, a y-axis direction, and a z-axis direction; a second set of controls configured to select one of the set of magnification levels for the first magnification level, the second magnification level, or the third magnification level; and a visual indicator configured to display a progress of the slide scanner. In some examples, the user interface is as described in the context of FIGS. 40A and 40B.

[0309] Solution 38. The system of solution 37, wherein the visual indicator comprises a progress bar.

[0310] Solution 39. The system of solution 37, wherein the first set of controls comprises arrow buttons that control the movement of the slide scanner in different directions, e.g., up, down, left, right, forward, backward, as shown in FIG. 40A.

[0311] In this document, the term "cellular sample" refers to any biological sample containing cells. Cellular samples can be a tissue sample or samples (e.g., any collection of cells) removed from a subject. The tissue sample can be a collection of interconnected cells that perform a similar function within an organism. A cellular sample can also be any solid or fluid sample obtained from, excreted by, or secreted by any living organism, including, without limitation, single-celled organisms, such as bacteria, yeast, protozoans, and amebae, multicellular organisms (such as plants or animals, including samples from a healthy or apparently healthy human subject or a human patient affected by a condition or disease to be diagnosed or investigated, such as cancer). In some embodiments, a cellular sample is mountable on a microscopeslide and includes, without limitation, a section of tissue, an organ, a tumor section, a smear, a frozen section, a cytology prep, or cell lines. An incisional biopsy, a core biopsy, an excisional biopsy, a needle aspiration biopsy (e.g., fine-needle aspiration (FNA)), a core needle biopsy, a stereotactic biopsy, an open biopsy, or a surgical biopsy can be used to obtain the sample.

[0312] The detailed descriptions of embodiments of the technology are not intended to be exhaustive or to limit the technology to the precise form disclosed above. Although specific embodiments of, and examples for, the technology are described above for illustrative purposes, various equivalent modifications are possible within the scope of the technology, as those skilled in the relevant art will recognize. The various embodiments described herein may also be combined to provide further embodiments.

[0313] From the foregoing, it will be appreciated that specific embodiments of the technology have been described herein for purposes of illustration, but well-known structures and functions have not been shown or described in detail to avoid unnecessarily obscuring the description of the embodiments of the technology.

[0314] It will also be appreciated that specific embodiments have been described herein for purposes of illustration, but that various modifications may be made without deviating from the technology. Further, while advantages associated with certain embodiments of the technology have been described in the context of those embodiments, other embodiments may also exhibit such advantages, and not all embodiments need necessarily exhibit such advantages to fall within the scope of the technology. Accordingly, the disclosure and associated technology can encompass other embodiments not expressly shown or described herein.

Claims

WHAT IS CLAIMED IS:1 . An apparatus for depositing, staining, and analyzing a cellular sample, comprising: a slide processing module, positioned at an upper front position of the apparatus, comprising: a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample, and a clamping module positioned at a lower front of the apparatus; an auxiliary systems module, positioned at a rear of the apparatus, comprising: a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol- based fixative bottle, and an electronics subsystem configured to execute, based on the input from the user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on a surface of a substrate; and an optical imaging and analysis system, configured to analyze the stained cellular sample, comprising: a microscopy subsystem configured to generate a digitized representation of the surface of the substrate at one or more magnification levels, and one or more processors configured to identify one or more portions of the digitized representation that include the stained cellular sample.

2. The apparatus of claim 1 , wherein the microscopy subsystem comprises a microscope that operates independently of the slide processing module.

3. The apparatus of claim 2, wherein the microscope is located adjacent to and separate from the apparatus.

4. The apparatus of claim 2, wherein the optical imaging and analysis system supports telepathology.

5. The apparatus of claim 1 , wherein the optical imaging and analysis system is positioned below the slide processing module.

6. The apparatus of claim 1 , wherein the optical imaging and analysis system is positioned above the slide processing module.

7. The apparatus of any of claims 1 to 6, wherein the optical imaging and analysis system is configured to support bright field, dark field, phase contrast, and fluorescent microscopy imaging modes.

8. The apparatus of any of claims 1 to 6, wherein the one or more processors is configured to: perform at least one pre-processing technique on the digitized representation.

9. The apparatus of claim 8, wherein the at least one pre-processing technique comprises a thresholding technique, an opening and closing technique, a surface roughness technique, a random orientation technique, an image sharpening technique, or a contrast normalization technique.

10. The apparatus of claim 8, wherein the one or more processors is further configured to perform, subsequent to the at least one pre-processing technique, a machine-learning (ML) classification algorithm on the digitized representation.11 . The apparatus of claim 10, wherein performing the ML classification algorithm comprises identifying the one or more portions of the digitized representation.

12. The apparatus of claim 10, wherein the ML classification algorithm is configured to detect diagnostic epithelial cells amidst red blood cells, artifacts, and glass slides.

13. The apparatus of any of claims 1 to 6, wherein the microscopy subsystem comprises a microscope and a slide scanner, each operable at the one or more magnification levels.

14. The apparatus of claim 13, wherein the optical imaging and analysis system is configured, as part of analyzing the stained cellular sample, to: generate, using the microscope at a first magnification level, a macroscopic view of the surface of the substrate;identify, using the one or more processors, coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection; generate, using the slide scanner at the first magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view; identify, using the one or more processors, one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection; and generate, using the slide scanner at a second magnification level higher than the first magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile.

15. The apparatus of claim 14, wherein the one or more tiles being identified begins prior to the first digitized representation being generated completes.

16. The apparatus of claim 14, wherein the user request is from a remote user.

17. The apparatus of claim 14, wherein an indication of whether the substrate comprising the stained cellular sample warrants further processing is based on a number of the one or more tiles.

18. A method of analyzing a cellular sample on a substrate, comprising: generating, using a microscope at a first magnification level, a macroscopic view of a surface of the substrate; identifying, using one or more processors, coordinates of a portion of the macroscopic view of the surface of the substrate that includes samples of interest with a first probability of detection; generating, using a slide scanner at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view; identifying, using the one or more processors, one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection; generating, using the slide scanner at a third magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a userrequest comprising an identifier of the at least one tile; and identifying, based on the second digitized representation, at least one cell type within the samples of interest.

19. The method of claim 18, wherein the one or more processors, the microscope, and the slide scanner are components of a cytopathology imaging and analysis unit.

20. The method of claim 18, wherein the first probability of detection is configured based on a first set of training data for a specimen region identification algorithm, and wherein the second probability of detection is configured based on a second set of training data for a relevant tile identification algorithm.21 . The method of claim 20, wherein the first probability of detection is greater than or equal to the second probability of detection.

22. The method of claim 20, wherein the first probability of detection is less than or equal to the second probability of detection.

23. The method of claim 18, wherein the cellular sample on the substrate is produced using a benchtop instrument for deposition and staining.

24. The method of claim 23, wherein the benchtop instrument comprises: a slide processing module, positioned at an upper front position of the benchtop instrument, comprising: a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample, and a clamping module positioned at a lower front of the benchtop instrument; and an auxiliary systems module, positioned at a rear of the benchtop instrument, comprising: a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol- based fixative bottle, and an electronics subsystem configured to execute, based on the input fromthe user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on the surface of the substrate.

25. The method of claim 24, wherein the cellular sample is in a transport solution, and wherein the deposition operation comprises spraying the cellular sample in the transport solution from a spray nozzle onto the surface of the substrate.

26. The method of claim 25, wherein identifying the coordinates of the portion of the macroscopic view or identifying the one or more tiles in the first digitized representation is based on at least one of a design of the spray nozzle, a distance between the spray nozzle and the surface of the substrate, a cell type or a cell concentration of the cellular sample, or a characteristic of the transport solution.

27. The method of claim 18, wherein the first magnification level is identical to the second magnification level.

28. The method of claim 18 or 27, wherein the third magnification level is greater than the first magnification level and the second magnification level.

29. The method of claim 18 or 27, wherein the third magnification level is identical to the second magnification level.

30. The method of claim 18, wherein the at least one cell type comprises a benign cell type, a malignant cell type, an atypical cell type, or an infectious cell type.31 . The method of claim 18, wherein the at least one cell type comprises a red blood cell type or a diagnostic epithelial cell type.

32. A system for analyzing a cellular sample on a substrate, comprising: a microscope configured to generate a macroscopic view of a surface of the substrate at a first magnification level of a set of magnification levels; a slide scanner, coupled to the microscope, configured to generate a digitized view of the surface of the substrate at each of the set of magnification levels; and one or more processors, coupled to the microscope and the slide scanner, configured to identify coordinates of a portion of the macroscopic view of the surface ofthe substrate that includes samples of interest with a first probability of detection; wherein the slide scanner is further configured to generate, at a second magnification level based on the coordinates, a first digitized representation of the portion of the macroscopic view, wherein the one or more processors is further configured to identify one or more tiles in the first digitized representation that include the samples of interest with a second probability of detection, wherein the slide scanner is further configured to generate, at a third magnification level, a second digitized representation of at least one tile of the one or more tiles in response to a user request comprising an identifier of the at least one tile, and wherein the one or more processors is further configured to identify, based on the second digitized representation, at least one cell type within the samples of interest.

33. The system of claim 32, wherein the first magnification level is identical to the second magnification level and the third magnification level.

34. The system of claim 32, wherein the third magnification level is greater than the first magnification level and the second magnification level.

35. The system of claim 32, wherein the at least one cell type comprises a benign cell type, a malignant cell type, an atypical cell type, an infectious cell type, a red blood cell type, or a diagnostic epithelial cell type.

36. The system of any of claims 32 to 35, further comprising: a benchtop instrument, comprising: a slide processing module, positioned at an upper front position of the benchtop instrument, comprising: a user interface configured to receive an input from a user that configures a deposition operation and a staining operation for the cellular sample, and a clamping module positioned at a lower front of the benchtop instrument; and an auxiliary systems module, positioned at a rear of the benchtop instrument, comprising:a chassis including a removable storage container configured to hold a plurality of reagent bottles, a buffer solution bottle, and an ethanol-based fixative bottle, and an electronics subsystem configured to execute, based on the input from the user, a pre-programmed protocol for the deposition operation and the staining operation, thereby generating a stained cellular sample on the surface of the substrate.

37. The system of any of claims 32 to 35, comprising: a user interface comprising: a first portion configured to display a digitized representation of the macroscopic view of the surface of the substrate; a second portion configured to display the second digitized representation of the at least one tile; a first set of controls configured to control a movement of the slide scanner in an x-axis direction, a y-axis direction, and a z-axis direction; a second set of controls configured to select one of the set of magnification levels for the first magnification level, the second magnification level, or the third magnification level; and a visual indicator configured to display a progress of the slide scanner.

38. The system of claim 37, wherein the visual indicator comprises a progress bar.

39. The system of claim 37, wherein the first set of controls comprises arrow buttons that control the movement of the slide scanner in different directions.

Images