Systems and methos for autofluorescence image analysis of in VIVO tissue

Abstract

Methods and systems for analyzing in vivo tissue utilizing autofluorescence signal to determine an anatomical structure and optionally a disease condition. The autofluorescence signal may be used to identify and segment regions of interest in the in vivo tissue. The autofluorescence signal may also be analyzed to determine entropy values of one or more regions of interest. The methods and systems may be used to aid a user during a surgical procedure.

Description

SYSTEMS AND METROS FOR AUTOFLUORESCENCE IMAGE ANALYSIS OF IN VIVO TISSUECross Reference to Related Applications

[0001] The present application is based on, claims priority to, and incorporates herein by reference in its entirety for all purposes, US Provisional Application Serial No. 63 / 671,898, fded July 16, 2024.Background

[0002] Primary hyperparathyroidism (PHPT) is a relatively common endocrine disorder of the parathyroid glands (PTGs) with approximately 100,000 new cases diagnosed each year in the US. Preoperative localization is important for surgical planning to inform laterality of the diseased gland(s) and facilitate intraoperative identification. However, localization accuracies range from 59% to 88%, resulting in a reported operative failure rate ranging from 2% to 15%. While refinement in imaging modalities and surgical approaches has led to improved patient outcomes, intraoperative localization of PTGs can remain a challenge for even the most experienced surgeons. Intraoperative techniques to confirm operative success, such as frozen section histological analysis and rapid parathyroid hormone (PTH) assay, are routinely utilized; however, they are time-consuming, costly, and subject to limited accuracy. Furthermore, unnecessary and inadvertent parathyroidectomy can have deleterious health implication on a patient and their ability to regulate calcium levels.

[0003] Fluorescence spectroscopy for intraoperative PTG identification has gained interest, however, these techniques may be associated with injection-related complications, allergic response, and radiation exposure, and have limited use in parathyroidectomy, as they are unable to differentiate hypercellular from normocellular glands.

[0004] Therefore, there is a need to develop systems and methods for in situ distinction of normocellular and hypercellular PTGs to avoid incorrect or unnecessary parathyroidectomy.Summary

[0005] The present disclosure provides systems and methods that overcome the aforementioned drawbacks by providing systems and methods for determining parathyroid pathological conditions from imaging data to assist a surgeon in surgical planning. The methods and systemscan quantify the differences in the autofluorescence pattern between normocellular PTGs, hypercellular PTGs, and thyroid tissue.

[0006] In one aspect of the present disclosure, a method for analyzing in vivo tissue is disclosed. The method comprises (a) receiving, via a processor, autofluorescence imaging data acquired from in vivo tissue of a subject, (b) identifying, via the processor, one or more regions of interest in the autofluorescence imaging data, (c) segmenting, via the processor, the one or more regions of interest in the autofluorescence imaging data to determine boundaries of the one or more regions of interest, (d) analyzing, via the processor, the autofluorescence imaging data to determine an entropy value of the one or more regions of interest, (e) determining, via the processor, an identity of the in vivo tissue of the subject for each of the one or more regions of interest, and (f) displaying, via a display, a report including an output label based on the identity of the in vivo tissue for each of the one or more regions of interest.

[0007] In another aspect of the present disclosure, a system for determining an identity of in vivo tissue of a subject, the system comprises a processor configured to (a) receive autofluorescence imaging data acquired from in vivo tissue of a subject, (b) identify one or more regions of interest in the autofluorescence imaging data, (c) segment the one or more regions of interest in the autofluorescence imaging data to determine boundaries of the one or more regions of interest, (d) analyze the autofluorescence imaging data to determine an entropy value of the one or more regions of interest, and (e) determine an identity of the in vivo tissue of the subject for each of the one or more regions of interest. The processor further comprises a user interface configured to display a report including an output label based on the identity of the in vivo tissue for each of the one or more regions of interest.

[0008] These aspects are non-limiting. Other aspects and features of the systems and methods described herein will be provided below.Brief Description of the Drawings

[0009] The foregoing features of embodiments will be more readily understood by reference to the following detailed description, taken with reference to the accompanying drawings, in which:

[0010] FIG. l is a block diagram of an example system for in vivo tissue identification, according to aspects of the present disclosure.

[0011] FIG. 2 is a block diagram of example components that can implement the system for in vivo tissue identification of FIG.1.

[0012] FIG. 3 is a schematic of an example system for in situ determination of parathyroid pathological conditions, according to aspects of the present disclosure.

[0013] FIG. 4 is an example graphical layout of the user interface display, according to aspects of the present disclosure.

[0014] FIG. 5 is a non-limiting example method for analyzing in vivo tissue, according to aspects of the present disclosure.

[0015] FIG. 6 is a scatter plot of the mean AFI, normalized AFI, and entropy and their performance per tissue type. Normalized AFI = tissue AFI / b ackground AFI.

[0016] FIG. 7 is a receiver operating characteristic (ROC) curve for entropy measurements of hypercellular parathyroid glands versus normocellular thyroid glands. AUC: Area under curve.

[0017] FIG. 8A is an image of a parathyroid gland in the visual spectrum.

[0018] FIG. 8B is an image of the parathyroid gland of FIG. 8A with a homogeneous autofluorescence pattern taken with a near infrared autofluorescence (NIRAF) camera.

[0019] FIG. 8C is a pixel-intensity image analysis of the parathyroid gland with a homogeneous autofluorescence pattern if FIGS. 8A-8B.

[0020] FIG. 9 is a flowchart illustrating the study exclusion parameters. After applying exclusion criteria, 376 images (including in-vivo and ex-vivo) from a total of 330 PTGs (273 normocellular and 57 hypercellular PTGs) were included in the analysis.

[0021] FIG. 10A shows box plots comparing in-vivo and ex-vivo mean AFI and AFI ratio data for target PTGs (N=46). fp<0.001.

[0022] FIG. 10B shows box plots comparing in-vivo and ex-vivo percent coefficient of variation (%CV) and entropy data for target PTGs (N=46). fp<0.001.

[0023] FIG. 11 is a plot of area under the ROC Curve (AUC) testing differences between PTG size only and a combination of PTG size and entropy. AUC results and the associated 95% CI are shown.Detailed Description

[0024] The discovery of natural PTG autofluorescence (AF) properties when excited with nearinfrared (NIR) light has led to widespread use of fluorescence spectroscopy in thyroid andparathyroid surgery. Mean AF intensity (AFI) of parathyroid tissue is reportedly 2-18 times greater than surrounding tissues (i.e., thyroid, lymph nodes, adipose tissue) and confers exceedingly high accuracy in concordance with frozen tissue histological diagnosis. Application of NIRAF during parathyroidectomy, however, has been limited thus far, likely owing to the presumption that healthy and diseased glands are poorly distinguished using NIRAF technology. However, the present disclosure provides systems and methods that can use autofluorescence patterns to differentiate tissue and physiological or pathological conditions of the tissue, such as between normocellular PTGs, hypercellular PTGs, and thyroid tissue. Further, the present disclosure provides systems and methods for determining parathyroid pathological conditions from such data.

[0025] Referring now to FIG. 1, an example of a system 100 for identifying an in vivo tissue of a patient from clinical data in accordance with some embodiments of the systems and methods described in the present disclosure is shown. As shown in FIG. 1, a computing device 150 can receive one or more types of imaging data (e.g., near infrared autofluorescence) from data source 102. In some embodiments, the computing device 150 can execute at least a portion of in vivo tissue identification system 104 to generate a report of the pathological condition of one or more PTGs from the clinical data source 102. In a non-limiting example, the report may identify one or more regions of interest as thyroid glands or PTGs. In another example, the report may identify normal (e.g., normocellular) PTGs or abnormal (e.g., hypercellular) PTGs. In an alternative example, the report may identify one or more PTGs as normal PTG, adenoma, or hyperplasia.

[0026] Additionally or alternatively, in some embodiments, the computing device 150 can communicate information about data received from the data source 102 to a server 152 over a communication network 154, which can execute at least a portion of the in vivo tissue identification system 104. In such embodiments, the server 152 can return information to the computing device 150 (and / or any other suitable computing device) indicative of an output of the in vivo tissue identification system 104.

[0027] In some embodiments, computing device 150 and / or server 152 can be any suitable computing device or combination of devices, such as a desktop computer, a laptop computer, a smartphone, a tablet computer, a wearable computer, a server computer, a virtual machine being executed by a physical computing device, and so on.

[0028] In some embodiments, data source 102 can be any suitable source of clinical data, such as another computing device (e.g., a server storing clinical data). In some embodiments, data source 102 can be local to computing device 150. For example, data source 102 can be incorporated with computing device 150. As another example, data source 102 can be connected to computing device 150 by a cable, a direct wireless link, and so on, Additionally or alternatively, in some embodiments, data source 102 can be located locally and / or remotely from computing device 150, and can communicate data to computing device 150 (and / or server 152) via a communication network (e.g., communication network 154).

[0029] In some embodiments, communication network 154 can be any suitable communication network or combination of communication networks. For example, communication network 154 can include a Wi-Fi network (which can include one or more wireless routers, one or more switches, etc ), a peer-to-peer network (e.g., a Bluetooth network), a cellular network (e.g., a 3G network, a 4G network, etc., complying with any suitable standard, such as CDMA, GSM, LTE, LTE Advanced, WiMAX, etc.), a wired network, and so on. In some embodiments, communication network 154 can be a local area network, a wide area network, a public network (e.g., the Internet), a private or semi-private network (e.g., a corporate or university intranet), any other suitable type of network, or any suitable combination of networks. Communications links shown in FIG. 1 can each be any suitable communications link or combination of communications links, such as wired links, fiber optic links, Wi-Fi links, Bluetooth links, cellular links, and so on.

[0030] Referring now to FIG. 2, an example of hardware 200 that can be used to implement data source 102, computing device 150, and server 152 in accordance with some embodiments of the systems and methods described in the present disclosure is shown. As shown in FIG. 2, in some embodiments, computing device 150 can include a processor 202, a display 204, one or more inputs 206, one or more communication systems 208, and / or memory 210. In some embodiments, processor 202 can be any suitable hardware processor or combination of processors, such as a central processing unit (“CPU”), a graphics processing unit (“GPU”), and so on. In some embodiments, display 204 can include any suitable display devices, such as a computer monitor, a touchscreen, a television, and so on. In some embodiments, inputs 206 can include any suitable input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, and so on.

[0031] In some embodiments, communications systems 208 can include any suitable hardware, firmware, and / or software for communicating information over communication network 154 and / or any other suitable communication networks. For example, communications systems 208 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particular example, communications systems 208 can include hardware, firmware and / or software that can be used to establish a Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.

[0032] In some embodiments, memory 210 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 202 to present content using display 204, to communicate with server 152 via communications system(s) 208, and so on. Memory 210 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 210 can include RAM, ROM, EEPROM, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some embodiments, memory 210 can have encoded thereon, or otherwise stored therein, a computer program for controlling operation of computing device 150. In such embodiments, processor 202 can execute at least a portion of the computer program to present content (e.g., images, user interfaces, graphics, tables), receive content from server 152, transmit information to server 152, and so on.

[0033] In some embodiments, server 152 can include a processor 212, a display 214, one or more inputs 216, one or more communications systems 218, and / or memory 220. In some embodiments, processor 212 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on. In some embodiments, display 214 can include any suitable display devices, such as a computer monitor, a touchscreen, a television, and so on. In some embodiments, inputs 216 can include any suitable input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, and so on.

[0034] In some embodiments, communications systems 218 can include any suitable hardware, firmware, and / or software for communicating information over communication network 154 and / or any other suitable communication networks. For example, communications systems 218 can include one or more transceivers, one or more communication chips and / or chipsets, and so on. Tn a more particular example, communications systems 218 can include hardware, firmware and / or software that can be used to establish a connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.

[0035] In some embodiments, memory 220 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 212 to present content using display 214, to communicate with one or more computing devices 150, and so on. Memory 220 can include any suitable volatile memory, nonvolatile memory, storage, or any suitable combination thereof. For example, memory 220 can include RAM, ROM, EEPROM, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some embodiments, memory 220 can have encoded thereon a server program for controlling operation of server 152. In such embodiments, processor 212 can execute at least a portion of the server program to transmit information and / or content (e g., data, images, a user interface) to one or more computing devices 150, receive information and / or content from one or more computing devices 150, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone), and so on.

[0036] In some embodiments, data source 102 can include a processor 222, one or more data acquisition system(s) 224, one or more communications systems 226, and / or memory 228. In some embodiments, processor 222 can be any suitable hardware processor or combination of processors, such as a CPU, a GPU, and so on.

[0037] Note that, although not shown, data source 102 can include any suitable inputs and / or outputs. For example, data source 102 can include input devices and / or sensors that can be used to receive user input, such as a keyboard, a mouse, a touchscreen, a microphone, a trackpad, a trackball, a stylus, a wand, and so on. As another example, data source 102 can include any suitable display devices, such as a computer monitor, a touchscreen, a television, etc., one or more speakers, and so on.

[0038] In some embodiments, communications systems 226 can include any suitable hardware, firmware, and / or software for communicating information to computing device 150 (and, in some embodiments, over communication network 154 and / or any other suitable communication networks). For example, communications systems 226 can include one or more transceivers, one or more communication chips and / or chip sets, and so on. In a more particularexample, communications systems 226 can include hardware, firmware and / or software that can be used to establish a wired connection using any suitable port and / or communication standard (e.g., VGA, DVI video, USB, RS-232, etc.), Wi-Fi connection, a Bluetooth connection, a cellular connection, an Ethernet connection, and so on.

[0039] In some embodiments, memory 228 can include any suitable storage device or devices that can be used to store instructions, values, data, or the like, that can be used, for example, by processor 222 to control the one or more data acquisition system(s) 224, and / or receive data from the one or more data acquisition system(s) 224; to images from data; present content (e.g., images, a user interface) using a display; communicate with one or more computing devices 150; and so on. Memory 228 can include any suitable volatile memory, non-volatile memory, storage, or any suitable combination thereof. For example, memory 228 can include RAM, ROM, EEPROM, one or more flash drives, one or more hard disks, one or more solid state drives, one or more optical drives, and so on. In some embodiments, memory 228 can have encoded thereon, or otherwise stored therein, a program for controlling operation of data source 102. In such embodiments, processor 222 can execute at least a portion of the program to generate images, transmit information and / or content (e.g., data, images) to one or more computing devices 150, receive information and / or content from one or more computing devices 150, receive instructions from one or more devices (e.g., a personal computer, a laptop computer, a tablet computer, a smartphone, etc.), and so on.

[0040] Referring now to FIG. 3, a system 300 for in situ analysis of, for example, parathyroid condition is shown. In a non-limiting example, the system 300 may be employed in a surgical setting to provide information to a surgeon (not shown) during a, for example, parathy roi dectomy .

[0041] An imaging system 302 may include a light source 304 and a detector 306 to acquire imaging data of a subject 308. In a non-limiting example, the imaging system 302 may acquire imaging data of the subject’s neck region 310. The neck region 310 may have an incision to expose the underlying tissues and structures, including, but not limited to, thyroid glands, PTGs, muscle, and airway structures.

[0042] In a non-limiting example, the light source 304 may a beam of light 312 in the nearinfrared (NIR) spectrum. In one, non-limiting example, the spectrum may span from 780 nm to 2500 nm. In one embodiment, the light 312 has a wavelength between 785 nm. The presentdisclosure recognizes that PTGs exhibit NIR autofluorescence at around 820-830 nm. Thus, the selected spectrum may be calibrated relative to a desired autofluorescence range or other mechanism for data acquisition. The autofluorescence light 314 is detected by detector 306.

[0043] In a non-limiting example, the imaging system 302 may be any device capable of emitting a wavelength in the NIR spectrum and detect the generated autofluorescence from the structures irradiated by the NIR light. The device may include components integrated into a single structure. Alternatively, the device may be comprised of separate components.

[0044] In an alternative configuration (not shown), full or partial samples of the thyroid, PTGs, or other structures of the neck region 310 may be excised and imaged by the imaging system 302 ex vivo. Furthermore, the samples may be frozen section samples.

[0045] The system 300 further includes a processor 316, configured to receive the imaging data acquired by the imaging system 302. Alternatively, the processor 316 may receive previously acquired imaging data from a storage system. In a non-limiting example, the processor 316 includes a plurality of processing units (physical or virtual) 318-324 for processing the imaging data. For example, an identification unit 318 may identify one or more regions of interest (RO I) in the imaging data. In one non-limiting example, the one or more ROI may be identified by the signal intensity variation between the one or more objects in the imaging data. For example, the PTGs may be identified in the imaging data based on their increased signal intensity. The present disclosure recognizes that the autofluorescence signal intensity is 2-18 times greater relative to the autofluorescence signal surrounding tissues (e.g., thyroid, lymph nodes, adipose tissue).

[0046] In a non-limiting example, the identification unit 318 may identify the one or more ROI based on a threshold value of the signal intensity. For example, objects in the imaging data with signal intensities above the threshold value may indicate an ROI, while objects below the threshold value constitute other tissue or background.

[0047] In a non-limiting example, the identification unit 318 may set a threshold value based on identifying a background signal intensity from a background region of the imaging data and subtracting the background signal intensity from the signal intensity of the one or more regions of interest. The background region does not overlap with any of the one or more ROI.

[0048] The processor 316 may further include an ROI segmentation unit 320 for segmenting the one or more ROI in the imaging data identified by the identification unit 318. In a non-limiting example, the borders of the one or more regions of interest are automatically or manually drawn.In an alternative embodiment, the identification unit 318 and the segmentation unit 320 may be a single processing unit.

[0049] In a non-limiting example, the processor 316 includes an analysis unit 322 configured to analyze the imaging data to determine an entropy value of the one or more ROI. In a nonlimiting example, the entropy value may also be referred to as the texture value of each of the one or more ROI and measures the similarity in black-and-white intensity between pixels within a selected area.Further, the measurement unit 322 may measure an area of each of the one or more ROI and the longest diameter of each of the one or more ROI (referred to herein as the “size” of each of the one or more ROI).

[0050] The processor 316 may further include an identity determination unit 324 to identify the in vivo tissue of the subject for each of the one or more ROI. For example, the identity determination unit 324 may identify an anatomical structure of the one or more ROI. For example, the label generation unit 324 may label one of the ROI as one of thyroid tissue or PTG. Further, the identity determination unit may provide a condition of the anatomical structure. For example, the identity determination unit 324 may label one or more of the PTG as normocellular PTG or hypercellular PTG.

[0051] In a non-limiting example, the identity of the in vivo tissue may be based on a probability, p, of the identity and optionally the condition of the anatomical structure of each of the one or more ROI. As an example, the output label may provide a 0.71 probability that a ROI is normocellular PTG, a 0.21 probability the ROI is hypercellular PTG, and a 0.08 probability that the ROI is thyroid tissue.

[0052] The identity determination unit 324 may utilize an equation, such as one defined bywhere E is the entropy value and S is the size of an ROI of each of the one or more ROI.

[0053] Equation (1) represents one possible probability calculation with entropy value and size as the only parameters (including their associated weights). Other parameters may also be considered such as, but not limited to, mean autofluorescence signal intensity, normalized autofluorescence signal intensity, and percent coefficient of variation (%CV).

[0054] In a non-limiting example, the processor 316 may also include a report generation unit 326 for generating a report to be output to a user interface 328. The report may include an outputlabel based on the identity of the in vivo tissue for each of the one or more ROI determined by the identity determination unit 324. In an alternative example, the report may also include the imaging data, the signal intensity, the area, the entropy value, or any combination thereof for each of the one or more ROI. In a non-limiting example, the report may also include a treatment or therapy suggestion based on identity of the in vivo tissue. For example, the report may suggest to a user or surgeon to perform a full or partial parathyroidectomy on one or more PTGs, avoid parathyroidectomy of the one or more PTGs, prescribe a pharmaceutical therapy, pursue no treatment, or the like.

[0055] In a non-limiting example, the report may include a summary indicator based on any combination of the output label, entropy value, size, area, signal intensity, etc. for a user or surgeon to easily determine the condition of the one or more ROI and what course of action to pursue. In a non-limiting example, the summary indicator may be color-coded. For example, the display may show a green color scheme or icon indicating to a user to proceed with a parathyroidectomy of the one or more ROI. A yellow color scheme or icon may indicate that the one or more ROI is normocellular PTG and to perform a manual check of the processed information. A red color scheme or icon may indicate that the one or more ROI is normocellular or thyroid tissue and to avoid excision. This example use of a red-yellow-green color scheme is non-limiting and may indicate other information to a user. Alternatively, the summary indicator may be an audio signal (e g., an alarm) or some other visual indicator (e.g. pop-up window).

[0056] In a non-limiting example, the processor 316 may be further connected to a user interface 328, which may include a display 330 and an input device 332. The display 330 may display a report including the output label, the imaging data, the annotated imaging data including the identified and / or segmented one or more ROI, and the area, size, and / or entropy values of the one or more ROI, the summary indicator, or any combination thereof.

[0057] The input device 332, may include any of a variety of input devices, such as those provided above. In a non-limiting example, a surgeon or user may use the input device 332 to manually identify and segment the one or more ROI by annotating the imaging data. Additionally or alternatively, the surgeon or user may use the input device 332 to adjust the processing steps automated by the processor 316. For example, the user may use the input device 332 to adjust one or more boundaries of each of the one or more segmented ROI generated fromthe segmentation unit 320. In another example, the user may use the input device 332 to select the one or more ROI and / or background region.

[0058] FIG. 4 shows a non-limiting example of a graphical layout 400 of the display 330. In this example, the graphical layout includes an acquisition information field 402 which may include, but is not limited to, details about the subject and imaging acquisition parameters. The graphical layout 400 may also include a summary indicator field 404 to provide the summary indicator directed to the identity and condition of one or more ROI as described above. Further, the graphical layout 400 may include a field 406 showing the imaging data, while field 408 may show the annotated imaging data including the segmentations and / or output label. Graphical layout 400 may also include a measurement field 410 of the ROI signal intensities, entropy values, areas, sizes, mean AFI, normalized AFI, coefficient of variation, or any other measurement parameter. This particular graphical layout 400 is non-limiting, and may include additional fields, fewer fields, and different layouts of the fields.

[0059] Referring now to FIG. 5, an example method 500 for analyzing in vivo tissue is provided. In a non-limiting example, the method may be practiced using any of the systems provided herein, for example processor 316 in FIG. 3. At step 502, the processor receives autofluorescence imaging data from in vivo tissue of the subject. Thereafter at step 504, the processor identifies one or more ROI in the autofluorescence imaging data. At step 506 the processor further segments the one or more ROI in the autofluorescence imaging data to determine the boundaries of each of the one or more ROI. At step 508, the processor analyzes the autofluorescence imaging data to determine an an entropy value of the one or more ROI. As provided previously, the processor may also determine an area and size of each of the one or more ROI. At step 510, the processor may determine an identity of the in vivo tissue of the subject for each of the one or more ROI. Furthermore, at step 510, the pathological condition of the identified anatomical structure may also be determined. At step 512, the processor may communicate with a user interface including a display to display a report including an output label based on the identity of the in vivo tissue for each of the one or more ROI. As provided above, the report may further include the imaging data, the annotated imaging data including the identified and / or segmented one or more ROI, the area, size, and / or entropy values of the one or more ROI, or any combination thereof. In a non-limiting example, the report may include a summary indicator of the output label, entropy value, size, area, signal intensity, etc. for a user or surgeon to easilydetermine the pathological condition of the one or more ROI and what course of action to pursue, as described previously.

[0060] The following examples are intended to provide non-limiting examples and implementations of the systems and methods described herein.

[0061] Example 1

[0062] 1.1 Methods

[0063] 1.1.1 Study design and subjects

[0064] Ethical approval was obtained from the Institutional Review Board at Mass. Eye and Ear Infirmary / Mass General Brigham, Boston, MA. Total thyroidectomy cases were initially excluded from enrollment into this study (total thyroidectomy cases were enrolled in other studies for logistical reasons). At a later stage, however, all cases were included regardless of type of surgery into this study.

[0065] 1.1.2 Surgical Procedures

[0066] Procedures were performed under general anesthesia using standard thyroidectomy and parathyroidectomy incisions and with intraoperative nerve monitoring (I0NM). The surgical approach for parathyroid surgery is to perform unilateral exploration for localized single adenomas and bilateral exploration for suspected multigland disease. Once sufficient dissection occurred, PTG candidates were visualized using the image-based NIRAF system Fluobeam LX (Fluoptics, Grenoble, France). Excised thyroid specimens, if applicable, were also examined. American Head and Neck Society autofluorescent consensus statement and Manufacturer’s protocol were closely followed for use of the NIRAF camera. Operating room (OR) surgical lights were turned off and ceiling lights were left on. The Fluobeam LX camera was held approximately 8 centimeters (cm) away from the surgical field. All resected PTGs were sent for frozen section and reviewed for final pathology.

[0067] 1.1.3 Study Variables

[0068] Patient demographics examined included age, gender, body mass index, and surgical procedure. PTG variables studied included size, location, image captured in-vivo / ex-vivo, and quantification data from NIRAF imaging. Pathological diagnosis was classified as normocellular (normal PTG), hypercellular (abnormal PTG), and thyroid tissue.

[0069] 1.1.4 Quantifying Autofluorescence Intensity (AFI), Normalized Intensity & Entropy

[0070] Deidentified images of parathyroid and thyroid tissues were recorded and uploaded to Imaged software (U.S. National Institutes of Health, Bethesda, Maryland, USA) for measurement of AF characteristics. To calculate the AFI and normalized AFI, the following steps were followed in each image: first, the borders of the PTGs were traced and selected using Imaged software. If the case was clinically judged to be an adenoma with a clearly differentiated parathyroid cap, the borders of the adenoma were traced and selected, excluding the cap. Images were collected from an area containing the capsule, not from within the gland itself. In particular, for adenomas, it is critical to avoid breaking the capsule to prevent potential seeding. Consequently, the images necessarily included the capsule to maintain the integrity of the gland and ensure safe handling. The selected area was then used to measure the mean and standard deviation of the PTG AFI. Second, an area of the same size from the adjacent central neck soft tissue in the same image was traced to measure the mean and standard deviation of the background AFI for in vivo images. For ex-vivo images, an area of the same size was selected from the adjacent surgical drape to use as a background with the same distance between the camera and the imaged PTG, which is recommended in similar published protocols. Third, normalized AFI was calculated by dividing mean PTG AFI by mean background AFI. Entropy of the PTGs was measured as an additional parameter. Entropy is an imaging parameter commonly used in texture analysis for computer vision which measures the similarity in black- and-white intensity between pixels within a selected area. To calculate entropy, the Imaged plugin Texture Analyzer (v0.4) was used.

[0071] 1.1.5 Statistical Analysis

[0072] Statistical analysis was performed using SPSS version 23.0 (SPSS Inc, Chicago, IL). Unpaired t-tests and ANOVA were respectively used to determine two- and three-group differences between normocellular PTGs, hypercellular PTGs, and thyroid tissue for variables including AFI, normalized AFI and entropy. Receiver operating characteristic (ROC) curves were derived from measurements of imaging parameters between normocellular and hypercellular PTGs to generate an area under the curve (AUC). For all tests, P values <0.05 were considered significant except when Bonferroni adjustment was used.

[0073] 1.2 Results

[0074] A total of 141 normocellular PTGs, 54 hypercellular PTGs, and 103 thyroid specimens in 103 patients were recorded using Fluobeam and included in this study. Baseline characteristics of patient demographics and surgeries are detailed in Table 1.1.Table 1.1. Patient Demographics and Baseline characteristics. t N (%)Characteristic „ _ ' , „„Or Mean ± SDTotal Number of Patients 103Age at surgery in years 52.2±16.7(mean ± SD)Sex, n(%)Female 78 (75.7)Male 25 (24.2)JMI28.1±7.2(mean ± SD)Surgery, n(%)Hemithyroidectomy 70 (70.0)Parathyroidectomy 30 (29.1)Total thyroidectomy 3 (2.9)

[0075] 1.2.1 AFI and Normalized AFI

[0076] AFI and normalized AFI were captured for both in-vivo and ex-vivo specimens (Table 1.2, FIG. 6). No significant difference was found in AFI between normocellular and hypercellular PTGs (mean = 139.6 vs 152.4, P=0.2282). The signal was not uniform when comparing in-vivo versus ex-vivo images. In normocellular PTGs, AFI of ex-vivo specimens was significantly higher compared to in-vivo specimens AFI (mean = 173.7 vs 136.2, P=0.0012). In hypercellular PTGs, there was no statistically-significant difference between ex-vivo versus in-vivo specimens (158.7 vs 139.4, P=0.0771).Table 1.2. Autofluorescence Intensity (AFI), Normalized AFI and Entropy by the Tissue Type and Imaging Setting (In-vivo or Ex-vivo)Normocellular PTG Hypercellular PTG Thyroid

[0077] When AFI was normalized to the background, normocellular PTGs were found to have significantly higher normalized AFI compared to hypercellular PTGs (2.9 vs 2.0, P<0.001). Normalized AFI was consistently lower in ex-vivo normocellular PTGs and hypercellular PTGs compared to in-vivo PTGs (P<0.001).

[0078] 1.2.2 Entropy

[0079] Entropy was calculated for all normocellular PTGs, hypercellular PTGs, and thyroid specimens, and was compared in in-vivo versus ex-vivo imaging (Table 1.2, FIG. 6). Entropy was found to be different across normocellular PTGs, hypercellular PTGs, and thyroid tissue (P<0.0001). There was no difference demonstrated between in-vivo and ex-vivo specimens in normocellular PTGs, hypercellular PTGs or thyroid tissue (P= 0.9693, P= 0.2294, and P= 0.5291, respectively).

[0080] ROC curve was created for entropy measurements of hypercellular versus normocellular PTGs, which demonstrated an AUC of 0.931 (FIG. 7 and Table 1.3).Table 1.3. Possible Cut-off Values for Entropy and the Respective Sensitivity and SpecificityCutoff Sensitivity (%) Specificity (%)5.89 93.7 70.06.10 90.9 77.46.41 79.6 90.1

[0081] 1.3 Discussion

[0082] This is the first reported use of entropy (image texture analysis) to differentiate normal from hypercellular parathyroid tissue. This is a step further towards using near-infrared autofluorescence (NIRAF) detection cameras for both identification of PTG tissue and assessment of pathogenicity.

[0083] The perceived benefit of NIRAF technology for identification of abnormal PTGs during parathyroidectomy has been limited until recently. However, studies have now demonstrated that hyperfunctioning glands exhibit more heterogeneous and low-intensity AF patterns compared to normo-functioning glands, indicating that parathyroid function may be predicted based on AF characteristics alone. The purpose of this study was to further characterize AF patterns of hypercellular and normocellular PTGs using the Fluobeam LX image-based NIRAF system in order to assess its potential for prediction of hypercellularity via converting the image derived from the system to a metric that can differentiate by parathyroid tissue type.

[0084] The current study investigating 195 PTGs identified in 103 cases of NIRAF deviceassisted thyroid and parathyroid surgeries revealed that normalized AFI, and to a greater extent, entropy, were adequate in distinguishing hypercellular from normocellular PTGs consistently across in-vivo and ex-vivo specimens.

[0085] The analysis revealed that raw AFI- that is, without normalization using the background - could not consistently demonstrate statistically significant differences between normocellular and hypercellular glands when studied both in-vivo and ex-vivo. For example, AFI was greater ex-vivo compared to in-vivo for normocellular PTGs, but this not was not statistically significant for hypercellular PTGs. It is postulated that AFI is inadequate due to variability between measurements arising from differences in residual background light as well as camera angle.

[0086] Normalization of AFI was able to overcome this, which revealed consistently lower normalized AFI values in the hypercellular PTG group compared to the normocellular PTG group (difference of 0.9 units). This is in agreement with previous studies by Kose et al. (difference of 0.8 units) and Akbulut et al. (difference of 0.3 units). Discordantly, it was also found that ex-vivo PTGs exhibited lower normalized AFI compared to in-vivo PTGs, which is puzzling given that AF of PTGs does not rely on vascularity and the normalization calculation theoretically removes the aforementioned confounding variables. However, it is postulated that there might be some effect arising from the fact that the backgrounds used for in-vivo and ex-vivo images of the same PTG are different (surgical tissue in-vivo versus surgical drapes ex- vivo).

[0087] The entropy characteristics of NIRAF images of normocellular and hypercellular PTGs were then analyzed in order to better capture and characterize pixel-by-pixel heterogeneity (FIGS. 8A-8C). It was found that entropy was a much better differentiator of normocellular and hypercellular PTGs with an improved AUC. The difference in entropy persisted for both in-vivo and ex-vivo PTGs. Thus, entropy characteristics, when compared to normalized AFI, represent a more robust measure of true tissue level AF and allow more optimal tissue differentiation.

[0088] This analysis is consistent with the work of previous studies showing greater AF heterogeneity in hypercellular glands compared to normocellular glands. In contrast to the study by Kose et al. in which heterogeneity was assessed subjectively and subsequent stratification was binary (i.e., homogenous vs heterogenous), objectively quantifiable imaging signals were examined. Akbulut et al. objectively quantified heterogeneity by calculating the standard deviation of AF pixel intensity (i.e., heterogeneity constant, HI), however, this calculation is not a true measure of inter-pixel heterogeneity but rather a reflection of the overall variance of intensities. By contrast, entropy accounts for the intensity of individual pixels and their relationship to adjacent pixels. As described herein, it is believed entropy is a superior metric given its potential to identify microscopic patchy AF patterns characteristically seen in hypercellular PTGs.

[0089] Despite potential limitations of the metric used by Akbulut et al, the machine learning algorithm employed to predict PTG hyperfunction using HI alone was associated with an AUC of 0.940; this was further improved to 0.987 when accounting for PTG volume and AF intensity. Still, there may be some impracticalities with clinical application of this methodology, as PTG volume cannot be readily incorporated into current real-time image-based NIRAF systems. It is also worthwhile to note that machine-learning algorithms often overfit data sets, which limits generalizability. It is believed that entropy is a simpler measurement with comparable accuracy in prediction of PTG hypercellularity (AUC 0.931). An optimized algorithm for autonomous PTG identification and characterization should permit objective verification of the quality of images without reliance on the surgeon’s visual assessment of AF patterns.

[0090] Although using each patient as their own control may initially seem appealing as a potential research approach, it was not the approach chosen for data collection or analysis in thisstudy. Tn clinical practice, many centers are increasingly adopting limited exploration, where only the abnormal gland is targeted and excised, making direct comparisons with the normal gland impossible. Even in unilateral exploration, where double adenomas are present, comparing the abnormal and normal glands is not feasible. To maximize the applicability of these findings across multiple institutions in the future, the focus needs to be on developing a technology that can identify abnormal glands independently, without requiring comparison to the normal gland. Furthermore, bilateral exploration is rarely performed unless there is bilateral disease, so cases where both the abnormal and normal glands are detected in the same patient are limited.

[0091] 1.4 Conclusions

[0092] The retrospective observational study using NIRAF images during parathyroid and thyroid surgeries reveals characteristic differences in patterns of AF intensity, normalized AFT and entropy between hypercellular and normocellular PTGs. These findings point towards a potential role of image-based NIRAF systems in predicting PTG hypercellularity.

[0093] Example 2

[0094] 2.1 Methods

[0095] 2.1.1 Study design and subjects

[0096] This study was approved by the Institutional Review Board at Mass General Brigham, Boston, MA. Patients who underwent thyroidectomy or parathyroidectomy were retrospectively identified by three independent surgeons at Massachusetts Eye and Ear Infirmary, Boston, MA between October 2020 and August 2021 using Fluobeam LX (Fluoptics, Grenoble, France).

[0097] Data was collected in consecutive patients undergoing thyroid or parathyroid surgery during the study period. Exclusion criteria were: (1) PTGs in patients who did not undergo thyroidectomy or parathyroidectomy (e.g., patients who underwent central neck dissection only), and (2) tissues that were bright on NIRAF but could not be visually confirmed by the surgeon as parathyroid tissue. Fluobeam LX (Fluoptics, Grenoble, France, part of Getinge) was used intraoperatively at the conclusion of surgery for visual confirmation of PTGs and was not employed for surgical decision-making within this series. As described previously, the Fluobeam system was used with operating room (OR) ceiling lights on and overhead surgical lights off. Data for each PTG was collected electronically in an institutional database. A registration number was assigned to each PTG and linked to a separate database of patient demographics.Because this study was observational, surgeons did not perform clinically unnecessary dissection for the sole intent of collecting PTG or NIRAF data.

[0098] Consistent with institutional practice and the previous work of others, pathological diagnosis was classified as normocellular (normal PTG) or hypercellular (abnormal PTG). A pathologic distinction between adenoma and hyperplasia was not made because this distinction is known to be problematic and associated with a high rate of clinical error. The distinction cannot be accurately made even with experienced pathologic examination. In this observational study, PTGs that appeared visually normal were not removed and were classified as normal / normocellular without pathologic confirmation, consistent with routine clinical practice. For the analysis, PTGs were excluded for which size could not be measured due to insufficient dissection.

[0099] 2.1.2 Study variables

[0100] Patient demographics included age, gender, body mass index, diagnosis, surgical procedure, and surgical outcome for hyperparathyroidism (cured / persistent). Variables pertaining to PTGs included PTG location, frozen section status, pathological diagnosis (normocellular / hypercellular / other), size of PTG (longest diameter measured intraoperatively), in-vivo vs. ex-vivo image capture, and quantification data from NIRAF imaging.

[0101] 2.1.3 Quantitative assessment of NIRAF images

[0102] Images obtained intraoperatively by the NIRAF camera system were transferred to an encrypted computer where each image was analyzed using Image J software (which is an open- source program). The software is able to calculate the mean and standard deviation (SD) of AutoFluorescence intensity (AFT) for any given area of interest.

[0103] To assess PTG brightness, two parameters were evaluated: mean AFI (as discussed earlier) and AFI normalized ratio (calculated by the following equation: AFI normalized ratio = mean AFI of PTG / mean AFI of the background). Here, the background was selected to be muscle or fat in the case of an in-vivo image, or the surgical drape in the case of an ex-vivo image. AFI normalized ratio is used to standardize the quantitative assessment of the PTG relative to the background in which it is imaged, Both the PTG and the background are imaged at a standardized distance of 8 cm from the Fluobeam camera with the overhead surgical lights off and operating room lights on.

[0104] To assess PTG heterogeneity, two parameters were evaluated: the percent coefficient of variation (%CV), which is calculated by the following equation: %CV = [SD / mean AFI of PTG] x 100) and entropy. To calculate entropy, the Image J plugin Texture Analyzer (v0.4) was applied.

[0105] 2.1.4 Statistical analyses

[0106] Statistical analyses were conducted using STATA software version 15.0 (Stata, College Station, TX). Pearson's chi-squared test was used for the comparison of categorical variables. The Wilcoxon signed-rank test was used to compare in-vivo and ex-vivo data for the same PTGs. Simple and multiple logistic regression models were used to analyze correlations between variables and PTG status. Candidate parameters for simple logistic regression models were chosen based on the aim of evaluating four separate quantification parameters: 1- AFI, 2- AFI normalized ratio, 3 -%CV, and 4- entropy, in addition to target PTG size. For multiple logistic regression models, the variable selection was focused on clinical assessment (brightness and heterogeneity) and PTG size. Initially, the AFI normalized ratio was examined because background correction should theoretically reduce unwanted variation. Entropy was included in the model because texture analysis could distinguish objects with different AF patterns (e g., regular non-uniform pattern [such as regular latticed pattern], random pattern, and localized pattern [such as half-tone gradient]). Multiple logistic regression analyses were performed before simple logistic regression analyses. Receiver operating characteristic (ROC) curves comparing the performance of NIRAF image parameters were used to generate an area under the curve (AUC). For all procedures, P values <0.05 were considered significant.

[0107] 2.2 Results

[0108] Three-hundred and thirty PTGs (273 normocellular PTGs and 57 hypercellular PTGs) from 190 subjects were included in the study. Table 2.1 summarizes the baseline characteristics of patients enrolled in this study. Of the 273 normocellular glands, 258 in-vivo and 25 ex-vivo images were included in the analysis. Of the 57 hypercellular glands, 39 in-vivo and 54 ex-vivo images were included in the analysis. The flow of data included for analysis is presented in FIG. 9. All patients with hypercellular PTGs were diagnosed with PHPT without a family history of multiple endocrine neoplasia. As noted above, glands were classified as either hypercellular or normocellular based on pathologic assessment if excised, or as normocellular if they were judged confidently by a senior surgeon to be visually normal and were not excised.Table 2.1 Baseline characteristics of enrolled patients (n=190)Number (%)Characteristic orMedian (IQR)Age, years 55 (41, 66)GenderMale 54 (28)Female 136 (72)BMI, kg / m226.6 (22.6, 32.7)Parathyroidectomy 50 (26)Thyroidectomy 151 (79)Total thyroidectomy 18 (9)Hemithyroidectomy 110 (58)Completion thyroidectomy 23 (12)

[0109] 2.2.1 In-vivo versus ex-vivo: differences in quantification parameters

[0110] In a preliminary analysis, verification of whether the imaging setting data (in-vivo versus ex-vivo) could be pooled or should be analyzed separately was sought. Of the cohort of 330 PTGs, 46 PTGs had both in-vivo and ex-vivo images captured. FIG. 10A shows the two brightness quantification parameters and FIG. 10B shows the two heterogeneity quantification parameters assessed for in-vivo and ex-vivo images of these target PTGs.

[0111] With respect to mean AFI and AFI normalized ratio - parameters representing brightness - PTGs were significantly different in in-vivo versus ex-vivo images. Ex-vivo images had higher mean AFI than in-vivo images, whereas ex-vivo images had a lower AFI normalized ratio. With respect to %CV and entropy - parameters representing heterogeneity - PTGs were not significantly different between in-vivo and ex-vivo imaging. Hence, in-vivo and ex-vivo data were pooled together for the assessment of %CV and entropy but performed analyses for mean AFI and AFI normalized ratio separately for in-vivo and ex-vivo images.

[0112] 2.2.2 NIRAF quantification parameters to differentiate hypercellular from normocellular PTGs

[0113] Multiple logistic regression analyses showed that entropy and PTG size were significant predictors for distinguishing hypercellular from normocellular PTGs under both in-vivo and ex- vivo conditions (Table 2.2). Simple logistic regression analyses showed that AFI normalized ratio and entropy had higher pseudo-R2compared to mean AFI and %CV, respectively. The area under the ROC curve (AUC) for entropy was significantly higher than that of %CV (0.938 vs. 0.809, p for difference < 0.001) for pooled in-vivo and ex-vivo images.Table 2.2 Simple and multiple logistic regression models using quantified values of NIRAF imaging under in-vivo and ex-vivo conditions to identify normocellular versus hypercellular PTGs

[0114] 2.2.3 Intraoperative probability calculator of hypercellular PTGs

[0115] An artificial intelligence (Al) algorithm was sought to be developed that could be incorporated into a NIRAF device, applicable to both in-vivo and ex-vivo image capture. The data suggests that entropy and size are the optimal parameters for predicting the probability of hypercellularity in this context. This probability can be calculated using logistic regression (Table 2.3) and is denoted by the following equation:1P 1 g25.70— 3.23 xEntropy— 0.68xSize

[0116] Notably, the addition of PTG size to entropy increased the AUC from 0.904 to 0.974 (P < 0.001) (FIG. 11), suggesting that augmentation of camera-based NIRAF technology with texture analysis can predict PTG hypercellularity with a high degree of accuracy when the size of gland is also taken into account.Table 2.3 Multiple logistic regression models using in-vivo and ex-vivo data togetherMultiple logistic regressionPseudo-R2, % Odds Ratio Coefficient 95% CI P value

[0117] 2.2.4 Pilot testing of the clinical utility of the intraoperative probability calculator

[0118] The probability calculator was pilot-tested to examine its utility in avoiding inadvertent resection of normal PTGs. In the study population, 5 PTGs were resected due to intraoperative suspicion for abnormality / hypercellularity and were ultimately found to be normocellular on pathological examination (Table 2.4). Probabilistic prediction for each of these 5 resected normal PTGs using the proposed calculator was retrospectively performed. The calculator output for theprobability of hypercellularity was less than 0.1% in 3 cases, 4.0% in the 4thcase, and 50.1% in the 5thcase.Table 2.4 Probabilistic prediction of hypercellularity in PTGs in histologically proven normal PTGs. n=5 Entropy Size, mm Probabilistic prediction of hypercellular PTG, %Case l 1.346 6 <0.1Case 2 3 354 5 <0.1Case 3 2.882 8 <0.1Case 4 5.917 5 4.0Case 5 5.852 10 50.1

[0119] 2.3 Discussion

[0120] The results of this study demonstrate that NIRAF texture analysis (entropy) in combination with PTG size can differentiate between normal and hypercellular PTGs. The probability calculator developed is able to accomplish this with an extremely high degree of accuracy (AUC = 0.974).

[0121] Based on institutional experience, hypercellular PTGs characteristically appear dimmer and more heterogeneous than normal PTGs. Mean AFI and AH normalized ratio were used to quantify brightness; %CV and entropy were used to quantify heterogeneity. Studies have shown mixed results with respect to AFI patterns. One study found no difference in mean AFI between normal and hypercellular PTGs, though the number of normal PTGs was quite small (n=6). Another study of just five subjects using visual assessment found no difference. Falco et al. reported that parathyroid adenomas showed significantly higher mean AFI than normal PTGs, a finding that directly contradicts the hypothesis herein. These divergent findings might be explained by the use of older Fluobeam models in these studies (i.e. Fluobeam 800), PDE-NEO II [Hamamatsu Photonics, Hamamatsu, Japan], or other camera-based NIRAF devices) with OR ceiling lights off. In contrast, the newer Fluobeam LX model was used with OR ceiling lights on and surgical lights off. Additionally, the working distance between the object and camera differs between NIRAF devices, and the mean AFI is susceptible to change with alterations in distance. The 8cm camera to target distance was rigidly adhered to. Variations in the degree of PTGdissection may also impact the extent of brightness. In sum, mean AFI appears to be an impractical parameter for differentiating between normal and abnormal PTGs.

[0122] AFI normalized ratio has been used in some studies to reduce the influence of cameraholding distance variability and differences in background brightness. DiMarco et al. employed a qualitative AFI normalized ratio by comparing PTGs with background thyroid and categorizing AFI as low, medium, or high based on the surgeon’s impression. They found no significant difference between normal and abnormal PTGs. By contrast, McWade et al. found hypercellular PTGs had a significantly lower AFI normalized ratio than normal PTGs. Kose et al. also found that hypercellular PTGs had a lower AFI normalized ratio (quantitative assessment), higher rate of heterogeneous pattern (qualitative assessment), and larger PTG size compared to normally functioning PTGs. While this study showed the AFI normalized ratio to be a better predictor than mean AFI, its usefulness in differentiating normal from abnormal PTGs did not persist on multiple logistic regression analyses. Additionally, a discrepancy was found between in-vivo and ex-vivo AFI normalized ratios which could be explained by differences in background compositions (i.e., muscle, fat, or surgical drape).

[0123] While brightness has been inconsistently shown to differentiate abnormal from normal PTGs, studies examining heterogeneity as a predictor have had more consistent findings. As noted above, Kose et al. found a higher rate of heterogeneous patterns for hyperfunctioning PTGs imaged in-vivo. Demarchi et al. corroborated that finding in 23 parathyroid adenomas imaged ex-vivo, though there was no imaging of normal PTGs for comparison. Notably, in both of these studies, heterogeneity was assessed using qualitative parameters. Akbulut et al. performed a quantitative assessment using CV and AFI normalized ratio and found abnormal PTGs were dimmer and more heterogeneous than normal PTGs. The current study using entropy corroborates this finding, suggesting that heterogeneity analysis can facilitate differentiation between normal and abnormal PTGs. Additionally, results show that entropy is a more optimal parameter for distinguishing abnormal from normal PTGs.

[0124] 2.4 Conclusions

[0125] This retrospective observational study shows that NIRAF image texture analysis and PTG size together are sufficient to distinguish abnormal from normal PTGs. Consequently, camera-based NIRAF devices augmented with the suggested hypercellularity probabilitycalculator could represent the next generation of AF technology. This type of technology offers potential time- and cost-savings, though further study of real-time application is needed.

[0126] As used in this specification and the claims, the singular forms “a,” “an,” and “the” include plural forms unless the context clearly dictates otherwise.

[0127] As used herein, “about”, “approximately,” “substantially,” and “significantly” will be understood by persons of ordinary skill in the art and will vary to some extent on the context in which they are used. If there are uses of the term which are not clear to persons of ordinary skill in the art given the context in which it is used, “about” and “approximately” will mean up to plus or minus 10% of the particular term and “substantially” and “significantly” will mean more than plus or minus 10% of the particular term.

[0128] As used herein, the terms “include” and “including” have the same meaning as the terms “comprise” and “comprising.” The terms “comprise” and “comprising” should be interpreted as being “open” transitional terms that permit the inclusion of additional components further to those components recited in the claims. The terms “consist” and “consisting of’ should be interpreted as being “closed” transitional terms that do not permit the inclusion of additional components other than the components recited in the claims. The term “consisting essentially of’ should be interpreted to be partially closed and allowing the inclusion only of additional components that do not fundamentally alter the nature of the claimed subject matter.

[0129] The phrase “such as” should be interpreted as “for example, including.” Moreover, the use of any and all exemplary language, including but not limited to “such as”, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed.

[0130] Furthermore, in those instances where a convention analogous to “at least one of A, B and C, etc.” is used, in general such a construction is intended in the sense of one having ordinary skill in the art would understand the convention (e.g., “a system having at least one of A, B and C” would include but not be limited to systems that have A alone, B alone, C alone, A and B together, A and C together, B and C together, and / or A, B, and C together ). It will be further understood by those within the art that virtually any disjunctive word and / or phrase presenting two or more alternative terms, whether in the description or figures, should be understood to contemplate the possibilities of including one of the terms, either of the terms, orboth terms. For example, the phrase “A or B” will be understood to include the possibilities of “A” or “B” or “A and B.”

[0131] All language such as “up to,” “at least,” “greater than,” “less than,” and the like, include the number recited and refer to ranges which can subsequently be broken down into ranges and subranges. A range includes each individual member. Thus, for example, a group having 1-3 members refers to groups having 1, 2, or 3 members. Similarly, a group having 6 members refers to groups having 1, 2, 3, 4, or 6 members, and so forth.

[0132] The modal verb “may” refers to the preferred use or selection of one or more options or choices among the several described embodiments or features contained within the same. Where no options or choices are disclosed regarding a particular embodiment or feature contained in the same, the modal verb “may” refers to an affirmative act regarding how to make or use an aspect of a described embodiment or feature contained in the same, or a definitive decision to use a specific skill regarding a described embodiment or feature contained in the same. In this latter context, the modal verb “may” has the same meaning and connotation as the auxiliary verb “can.”

Claims

Claims1. A method for analyzing in vivo tissue, the method comprising:(a) receiving, via a processor, autofluorescence imaging data acquired from in vivo tissue of a subject,(b) identifying, via the processor, one or more regions of interest in the autofluorescence imaging data,(c) segmenting, via the processor, the one or more regions of interest in the autofluorescence imaging data to determine boundaries of the one or more regions of interest,(d) analyzing, via the processor, the autofluorescence imaging data to determine an entropy value of the one or more regions of interest,(e) determining, via the processor, an identity of the in vivo tissue of the subject for each of the one or more regions of interest, and(f) displaying, via a display, a report including an output label based on the identity of the in vivo tissue for each of the one or more regions of interest.

2. The method of claim 1, further comprising receiving a user-selection of the one or more regions of interest.

3. The method of claim 1, wherein identifying the one or more regions of interest is based on a threshold value of an autofluorescence signal intensity.

4. The method of claim 3, wherein the one or more regions of interest have a signal intensity above the threshold value.

5. The method of claim 3, further comprising setting the threshold value, via the processor, based on identifying a background signal intensity from a background region of the autofluorescence imaging data, and subtracting the background signal intensity from the autofluorescence signal intensity of the one or more regions of interest, wherein the background region does not overlap with the one or more regions of interest.

6. The method of claim 1, further comprising determining a probability of each of the one or more regions of interest being a thyroid gland, a normocellular parathyroid gland (PTG), or a hypercellular PTG to determine the identity of the in vivo tissue.

7. The method of claim 6, wherein the probability, / ?, is defined bywhere E is the entropy value of each of the one or more regions of interest and S is a longest diameter of each of the one or more regions of interest.

8. The method of claim 7, wherein the entropy value is a measure of a similarity in a black-and- white intensity between pixels of the one or more regions of interest.

9. The method of claim 1, further comprising automatically identifying and segmenting, via the processor, the one or more regions of interest.

10. The method of claim 1, further comprising receiving, via a user interface, user-identified segments indicating the one or more regions of interest.

11. The method of claim 1, further comprising receiving, via a user interface, user adjustments to the boundaries of the one or more regions of interest using a input device.

12. A system for determining an identity of in vivo tissue of a subject, the system comprising: a processor configured to:(a) receive autofluorescence imaging data acquired from in vivo tissue of a subject,(b) identify one or more regions of interest in the autofluorescence imaging data,(c) segment the one or more regions of interest in the autofluorescence imaging data to determine boundaries of the one or more regions of interest,(d) analyze the autofluorescence imaging data to determine an entropy value of the one or more regions of interest,(e) determine an identity of the in vivo tissue of the subject for each of the one or more regions of interest; and a user interface configured to display a report including an output label based on the identity of the in vivo tissue for each of the one or more regions of interest.

13. The system of claim 12, wherein the user interface is configured to receive a user-selection of the one or more regions of interest.

14. The system of claim 12, wherein the one or more regions of interest are identified based on a threshold value of an autofluorescence signal intensity.

15. The system of claim 14, wherein the one or more regions of interest have a signal intensity above the threshold value.

16. The system of claim 14, wherein the processor is further configured to set the threshold value based on identifying a background signal intensity from a background region of the autofluorescence imaging data, and subtracting the background signal intensity from the autofluorescence signal intensity of the one or more regions of interest, wherein the background region does not overlap with the one or more regions of interest.

17. The system of claim 12, wherein the processor is further configured to determine a probability of each of the one or more regions of interest being a thyroid gland, a normocellular parathyroid gland (PTG), or a hypercellular PTG to determine the identity of the in vivo tissue.

18. The system of claim 17, wherein the probability, / ?, is defined bywhere E is the entropy value of each of the one or more regions of interest and S is a longest diameter of each of the one or more regions of interest.

19. The system of claim 18, wherein the entropy value is a measure of a similarity in a black- and-white intensity between pixels of the one or more regions of interest.

20. The system of claim 12, wherein the processor is further configured to automatically identify and segment the one or more regions of interest.

21. The system of claim 12, wherein the processor is further configured to receive user-identified segments indicating the one or more regions of interest via the user interface.

22. The system of claim 12, wherein the processor is further configured to receive user adjustments to the boundaries of the one or more regions of interest.

Images