Method and system of assessing surgical margins of a post-neoadjuvant chemotherapy (NACT) breast cancer specimen

Abstract

According to embodiments of the present invention, a method of assessing margins of a post-Neoadjuvant Chemotherapy (NACT) breast cancer specimen is provided. The method includes obtaining real-time ultrasound images of the post-NACT breast cancer specimen; simultaneously performing photoacoustic imaging on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen; fused with the real-time ultrasound images, analyzing the spatial distributions of the plurality of chromophores along the tissue boundary of the post-NACT breast cancer specimen to determine margins; and classifying the margins into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer specimen. According to further embodiments, an apparatus for assessing margins of a post-NACT breast cancer specimen is also provided.

Description

METHOD AND SYSTEM OF ASSESSING SURGICAL MARGINS OF A POSTNEOADJUVANT CHEMOTHERAPY (NACT) BREAST CANCER SPECIMENCross-Reference To Related Application

[0001] This application claims the benefit of priority of Singapore patent application No. 10202402049P, filed 12 July 2024, the content of it being hereby incorporated by reference in its entirety for all purposes.Technical Field

[0002] Various embodiments relate to an apparatus and a method of assessing surgical margins of a post-Neoadjuvant Chemotherapy (NACT) breast cancer specimen.Background

[0003] Neoadjuvant chemotherapy (NACT) has become a mainstay in the treatment of locally advanced breast cancer It allows downstaging of tumours for breast-conserving surgery (BCS) with similar long-term oncologic outcomes as compared to mastectomy. However, this comes at a short-term expense of locoregional disease recurrence risk, which may be reduced by obtaining clear / negative margins during BCS (i.e. complete tumour excision). It is hence imperative to have an accurate intra-operative margin assessment tool during BCS for patients who underwent NACT.

[0004] Current mtra-operative margin assessment tools (e.g. clinical palpation or ultrasound) are often ineffective. The wide variety of unpredictable changes after NACT often results in heterogeneous changes of the breast stroma, resulting in indistinct borders on ultrasound (US). In addition, there may be scattered foci of residual tumour in the tumour bed that are too small to be distinguished on both US and clinical inspection / examination of the excised breast tissue. Frozen section analysis (FS A), the gold standard, is inaccurate in predicting false negatives in NACT specimens. Moreover, it is not practised in many institutions due to its high cost and long waiting-time.

[0005] Therefore, there is an unmet need for a precise intra-operative margin assessment tool in real-time, which addresses at least these obstacles for patients undergoing BCS after NACT to enhance surgical outcomes and patient care.Summary

[0006] According to an embodiment, a method of assessing margins of a post-Neoadjuvant Chemotherapy (NACT) breast cancer specimen is provided. The method includes obtaining real-time ultrasound images of the post-NACT breast cancer specimen; simultaneously (or substantially simultaneously) performing photoacoustic imaging on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen; fused with the real-time ultrasound images, analyzing the spatial distributions of the plurality of chromophores along the tissue boundary of the post-NACT breast cancer specimen to determine margins; and classifying the margins into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer specimen.

[0007] According to an embodiment, an apparatus for assessing margins of a post-NACT breast cancer specimen is provided. The apparatus includes an ultrasound device configured to capture real-time ultrasound images of the post-NACT breast cancer specimen; a photoacoustic imaging module configured to perform photoacoustic imaging on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen; and a computing module. The ultrasound device and the photoacoustic imaging module are simultaneously operable. The computing module is configured to analyze, when fused with the real-time ultrasound images, the spatial distributions of the plurality of chromophores along the tissue boundary of the post-NACT breast cancer specimen to determine margins; and classify the margins into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer specimen.Brief Description of the Drawings

[0008] Tn the drawings, like reference characters generally refer to like parts throughout the different views. The drawings are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of the invention. In the following description, various embodiments of the invention are described with reference to the following drawings, in which:

[0009] FIG. 1 shows a flow chart illustrating a method of assessing margins of a postNeoadjuvant Chemotherapy (NACT) breast cancer specimen, according to various embodiments.

[0010] FIG. 2 shows a schematic cross-sectional view of an apparatus for assessing margins of a post-NACT breast cancer specimen, according to various embodiments.

[0011] FIG. 3A shows a clinical in-vivo ultrasound image depicting presence of an irregular hypoechoic mass in a left breast of a patient (which was biopsied) proven to be a grade 3 invasive carcinoma, according to one example.

[0012] FIG. 3B shows an ex-vivo US image of the surgically excised specimen depicting a small residual hypoechoic mass which is in keeping with partial / incomplete response, according to the example in FIG. 3 A.

[0013] FIG. 3C shows a haematoxylin and eosin (H&E) stained microscopic image of specimen depicting residual tumour that is at least 3 mm away from all margins, according to the example of FIG. 3 A.

[0014] FIGS. 3D to 3F respectively show (left) the corresponding total lipid, hemoglobin, and collagen distribution maps generated from photoacoustic (PA) imaging, and (right) their inverses, according to the example of FIG 3 A.

[0015] FIG. 4A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 3.5 cm irregular hypoechoic mass in the left breast of a patient (which was biopsied) proven to be a grade 2 invasive carcinoma, according to one example.

[0016] FIG. 4B shows an ex-vivo US post-NACT image depicting no residual lesion upon surgical excision, highly suggestive of pathological complete response (pCR), according to the example of FIG 4 A.

[0017] FIG. 4C shows a H&E stained microscopic image of specimen depicting areas of fibrosis on low power view, with residual high-grade DC1S (ductal carcinoma m-situ) extending towards the posterior margins, according to the example of FIG. 4A

[0018] FIGS. 4D to 4F respectively show (left) the corresponding lipid, total hemoglobin (Hbt) and collagen distribution maps generated from PA imaging and (right) their inverses, according to the example of FIG. 4A.

[0019] FIG. 5A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 1.5 cm irregular hypoechoic mass in the left breast of a patent (which was biopsied) proven to be a grade 3 invasive carcinoma with associated ductal carcinoma in- situ, according to one example.

[0020] FIG 5B shows an ex-vivo US post-NACT image depicting presence of residual lesion upon surgical excision, in keeping with partial / incomplete response, according to the example of FIG. 5 A.

[0021] FIG. 5C shows a H&E stained microscopic image of specimen depicting residual tumour with some response to presurgical therapy, according to the example of FIG. 5 A.

[0022] FIGS. 5D to 5F respectively show (left) the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging, and (right) their inverses, according to the example of FIG. 5 A.

[0023] FIG. 6A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 2.0 cm irregular hypoechoic mass in the right breast of a patent (which was biopsied) proven to be a grade 3 invasive carcinoma, according to one example.

[0024] FIG. 6B shows an ex-vivo US post-NACT image depicting presence of residual lesion upon surgical excision, in keeping with partial / incomplete response, according to the example of FIG. 6A.

[0025] FIG 6C shows a H&E stained microscopic image of specimen depicting residual tumour that is away from the margins, with surrounding fibrosis (tumour bed) extending closer to the margins, according to the example of FIG. 6A.

[0026] FIGS. 6D to 6F respectively show (left) the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging, and (right) their inverses, according to the example of FIG. 6A.

[0027] FIG. 7 shows a flowchart depicting the summary of the overall findings for assessment, according to one example.

[0028] FIG. 8A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 2.0 cm irregular hypoechoic mass in the left breast of a patient (which was biopsied) proven to be a grade 3 invasive carcinoma, according to one example.

[0029] FIG. 8B shows an ex-vivo US post-NACT image depicting presence of residual mass upon surgical excision, in keeping with partial / incomplete response, according to the example of FIG. 8 A.

[0030] FIG. 8C shows a H&E stained microscopic image of specimen depicting residual tumour that is close to but not involving the posterior and inferior margins, according to the example of FIG 8 A.

[0031] FIGS. 8D to 8F respectively show (left) the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging, and (right) their inverses, according to the example of FIG. 8 A.

[0032] FIG. 9A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 3.0 cm lobulated heterogeneous mass in the left breast of a patient (which was biopsied) proven to be a grade 2 invasive carcinoma, according to one example.

[0033] FIG 9B shows an ex-vivo US post-NACT image depicting presence of residual mass upon surgical excision, in keeping with partial / incomplete response, according to the example of FIG. 9A.

[0034] FIG. 9C shows a H&E stained microscopic image of specimen depicting residual tumour that is close to but not involving the posterior and inferior margins, according to the example of FIG. 9A.

[0035] FIGS. 9D to 9F respectively show (left) the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging, and (right) their inverses, according to the example of FIG. 9A.

[0036] FIG. 10A shows a schematic view of a US-PA specimen imaging setup, according to an example.

[0037] FIG. 10B shows a photograph of a 2D array curvilinear US-OT probe used in the

[0038] FIG. 10C shows a photograph of a freshly excised breast cancer specimen with margins labelled using silk stitches under observation in the setup described in FIG. 10A.

[0039] FIG. 10D shows a schematic view of the specimen of FIG. 10C illustrating the different margins based on the silk thread in original positioning and the sections of ultrasound images acquired from medial to lateral margins. Top inset shows a schematic view of the specimen in a rotated position, while bottom inset shows a schematic view of the specimen in a flipped position.

[0040] FIG. 10E shows a normalized absorption spectra of endogenous chromophores in the excised breast specimen of FIG. 10C from the wavelength of 700 nm - 1100 nm based on which the acquired acoustic signals were unmixed to yield distribution maps of HbT (Hb+HbO2), collagen and lipidDetailed Description

[0041] The following detailed description refers to the accompanying drawings that show, by way of illustration, specific details and embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. Other embodiments may be utilized and structural, logical, and electrical changes may be made without departing from the scope of the invention. The various embodiments are not necessarily mutually exclusive, as some embodiments can be combined with one or more other embodiments to form new embodiments.

[0042] Embodiments described in the context of one of the methods or devices are analogously valid for the other methods or devices. Similarly, embodiments described in the context of a method are analogously valid for a device, and vice versa.

[0043] Features that are described in the context of an embodiment may correspondingly be applicable to the same or similar features in the other embodiments. Features that are described in the context of an embodiment may correspondingly be applicable to the other embodiments, even if not explicitly described in these other embodiments. Furthermore, additions and / or combinations and / or alternatives as described for a feature in the context of an embodiment may correspondingly be applicable to the same or similar feature in the other embodiments.

[0044] In the context of various embodiments, the articles “a”, “an” and “the” as used with regard to a feature or element include a reference to one or more of the features or elements.

[0045] In the context of various embodiments, the phrase “substantially” may include “exactly” and a reasonable variance.

[0046] In the context of various embodiments, the term “about” or “approximately” as applied to a numeric value encompasses the exact value and a reasonable variance.

[0047] As used herein, the term “and / or” includes any and all combinations of one or more of the associated listed items.

[0048] As used herein, the phrase of the form of “at least one of A or B” may include A or B or both A and B. Correspondingly, the phrase of the form of “at least one of A or B or C”, or including further listed items, may include any and all combinations of one or more of the associated listed items.

[0049] As used herein, the expression “configured to” may mean “constructed to” or “arranged to”.

[0050] Various embodiments relate to utility of photoacoustic patterns in intra-operative margin assessment of breast cancer post-Neoadjuvant Chemotherapy (NACT). Photoacoustic Tomography (PA) is a hybrid optical imaging modality, which provides endogenous differentiation ofthe major tissue components / contents such as lipids, collagen and hemoglobin which serve as biomarkers, e g. in breast tissue, in surgical margin assessment. When fused with ultrasound, PA potentially resolves the heterogeneous changes on ultrasound (US) by providing additional biochemical data (e.g. collagen from post-NACT collagenization and fibrosis) which may improve diagnostic confidence of the surgical margins.

[0051] FIG. 1 shows a flow chart illustrating a method 100 of assessing margins of a post- NACT breast cancer specimen, according to various embodiments. As seen in FIG. 1, at Step 102, real-time ultrasound images of the post-NACT breast cancer specimen are obtained. At Step 104, photoacoustic imaging is simultaneously (or substantially simultaneously) performed on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen. At Step 106, fused with the real-time ultrasound images, the spatial distributions of the plurality of chromophores are analyzed along the tissue boundary ofthe post-NACT breast cancer specimen to determine (or identify) margins. In other words, the obtainment of the real-time ultrasound images and the photoacoustic imaging may be carried out at about the same time, with negligible time difference or time lag. The spatial distributions of the plurality of chromophores at an instantaneous point in time may be fused with the ultrasound image(s) captured at that instantaneous point in time for the analysis. At Step 108, the margins are classified into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer specimen.

[0052] The margins may be interchangeably referred to as surgical margins.

[0053] By fusing the spatial distributions of the chromophores and the ultrasound images (as set out in Step 106), fusion imaging (or hybrid imaging) may be performed. Fusion imaging involves the combinations of images from different modalities to create a hybrid image.

[0054] The classified margins (from Step 108) may be positive margins if one of the two distinct situations is where cancer cells are present along at least part of the tissue boundary of the post-NACT breast cancer specimen. On the other hand, the classified margins (from Step 108) may be negative margins if the other of the two distinct situations is where cancer cells are absent along the tissue boundary of the post-NACT breast cancer specimen. The classified margins may be represented with binary scores based on the distinct situations.

[0055] The plurality of chromophores may include lipids, hemoglobin and collagen. Other chromophores may be, e.g. elastin, myoglobin or melanin.

[0056] In other words, the method 100 is a pioneering method of assessing the margins of post-NACT breast cancer specimens, where at its core, employs an innovative approach to evaluate margin status by thoroughly analyzing the distribution of chromophores along the margins of surgically excised specimens. Through advanced photoacoustic imaging techniques, the method 100 precisely maps the spatial distribution of key chromophores, such as lipid, hemoglobin and collagen (biochemical information) across the tissue boundary of NACT specimens, based on which, the method 100 classifies margins into distinct categories, including positive, and negative margins, providing surgeons with information that can be acted on during the surgical procedure. It should be appreciatedthat the method 100 uses ultrasound image to guide the imaging region to perform tumour margin detection and does not rely on markers or marker clips to mark the location(s) of tumour(s). Further, the method 100 is a PA and ultrasound hybrid imaging method for breast tumour applications and is designed to fit the purpose of neoadjuvant chemotherapy monitoring (with sufficient clinical data).

[0057] In various embodiments, if the spatial distribution of the lipids reflects disruption of a layer of the lipids, classifying the margins at Step 108 may include comparing a collagen signal intensity from the spatial distribution of the collagen with a signal intensity of a tumour bed of the post-NACT breast cancer specimen; if the collagen signal intensity at the margins is determined to be equal or higher than the signal intensity of the tumour bed, a first criterion is considered met; comparing a hemoglobin signal intensity of the spatial distribution of the hemoglobin with the signal intensity of the tumour bed or surroundings of the tumour bed, if the hemoglobin signal intensity at the margins is determined to be higher than the signal intensity of the tumour bed or the surroundings, a second criterion is considered met; if the collagen signal intensity or the hemoglobin signal intensity at the margins shows a direct continuity to the tumor bed or at least part of a tumour with (or comprising) the tumour bed, a third criterion is considered met; and classifying the margins based on at least one of the first criterion, the second criterion and the third criterion.

[0058] In the context of various embodiments, the expression “direct continuity” (or interchangeably, “direct connection”) may refer to vascularity, which is the visibility of veins, which may be influenced by blood pressure, muscle pressure, and fat levels.

[0059] For example, classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion may include for no residual lesion observed from the real-time ultrasound images, classifying the margins as positive margins if the spatial distribution of the lipids reflects disruption of a layer of the lipids and if all of the first criterion, the second criterion and the third criterion are considered met, or alternatively, for no residual lesion observed from the real-time ultrasound images, classifying the margins as negative margins if the spatial distribution of the lipids reflects disruption of the layer of the lipids and if none or only one or only two of the first criterion, the second criterion and the third criterion are considered met.

[0060] In this example, classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion may further include determining a presence of post-therapy changes at the margins, and if affirmative, the classified positive margins may present a likelihood of false positive margins Tn otherwise, if there are posttherapy changes at the margins, the classified positive margins may probably be false positive margins.

[0061] In a different example, classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion may include for residual lesion observed from the real-time ultrasound images, classifying the margins as positive margins if the spatial distribution of the lipids reflects disruption of a layer of the lipids and if only two or all of the first criterion, the second criterion and the third criterion are considered met; or alternatively, for residual lesion observed from the real-time ultrasound images, classifying the margins as negative margins if the spatial distribution of the lipids reflects disruption of the layer of the lipids and if none or only one of the first criterion, the second criterion and the third criterion are considered met.

[0062] In this different example, classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion may further include determining a presence of post-therapy changes at the margins and / or an extension of tumour cells into the margins, and if affirmative, the classified positive margins may present a likelihood of false positive margins. In otherwise, if there are post-therapy changes at the margins and / or there is an extension of tumour cells into the margins, the classified positive margins may probably be false positive margins. The extension may include a distance ranging from 1 mm to 2 mm.

[0063] In various embodiments, classifying the margins at Step 108 may include if the spatial distribution of the lipids reflects continuity or no disruption of a layer of the lipids, classifying the margins as negative margins.

[0064] The margins may include or may refer to superior margins, anterior margins, inferior margins, medial margins, lateral margins or any combinations thereof.

[0065] In various embodiments, the spatial distributions of the plurality of chromophores may be obtainable upto a depth of 5 mm of a specimen excised from the patient with the post-NACT breast cancer specimen during the breast-conserving surgery.

[0066] In other embodiments, the spatial distributions of the plurality of chromophores may be obtainable beyond a depth of 5 mm of a specimen excised from the patient with the post-NACT breast cancer specimen during the breast-conserving surgery, e g. the depth of about 6 mm, about 7 mm, about 8 mm, about 9 mm, about 10 mm, or up to a few centimeters.

[0067] While the method described above is illustrated and described as a series of steps or events, it will be appreciated that any ordering of such steps or events is not to be interpreted in a limiting sense. For example, some steps may occur in different orders and / or concurrently with other steps or events apart from those illustrated and / or described herein. In addition, not all illustrated steps may be required to implement one or more aspects or embodiments described herein Also, one or more of the steps depicted herein may be carried out in one or more separate acts and / or phases.

[0068] FIG. 2 shows a schematic cross-sectional view of an apparatus 220 (or system) for assessing margins of a post-NACT breast cancer specimen, according to various embodiments. As seen in FIG. 2, the apparatus 220 includes an ultrasound device 222 configured to capture real-time ultrasound images of the post-NACT breast cancer specimen; a photoacoustic imaging module 224 configured to perform photoacoustic imaging on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen; and a computing module 226. The ultrasound device 222 and the photoacoustic imaging module 224 are simultaneously (or substantially simultaneously) operable, and are in communication with the computing module 226, as denoted by lines 228, 230, respectively.

[0069] The computing module 226 is configured to analyze, when fused with the real-time ultrasound images, the spatial distributions of the plurality of chromophores along the tissue boundary of the post-NACT breast cancer specimen to determine (or identify) margins; and classify the margins into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer

[0070] The apparatus 220 may include the same or like elements or components as those of the method 100 of FIG. 1, and as such, the like elements may be as described in the context of the method 100 of FIG. 1, and therefore some corresponding descriptions are omitted here.

[0071] In various embodiments, the photoacoustic imaging module 224 may include a radio frequency source configured to deliver radio frequency pulses into the post-NACT breast cancer specimen; an ultrasonic transducer configured to detect ultrasonic waves generated by the post-NACT breast cancer specimen upon absorption of the radio frequency pulses and conversion into heat; and a data collector in communication with the ultrasonic transducer, wherein the data collector is configured to analyze the detected ultrasonic waves to produce images including the spatial distributions of the plurality of chromophores. For example, the photoacoustic imaging module 224 may further include an optical arrangement configured to direct and focus the radio frequency pulses into the post-NACT breast cancer specimen. For example, the radio frequency source may include a pulsed laser.

[0072] The plurality of chromophores may include lipids, hemoglobin and collagen. Other chromophores may be, e g. elastin, myoglobin or melanin.

[0073] If the spatial distribution of the lipids is determined to reflect disruption of a layer of the lipids, the computing module 226 may be configured to compare a collagen signal intensity from the spatial distribution of the collagen with a signal intensity of a tumour bed of the post-NACT breast cancer specimen; if the collagen signal intensity at the margins is determined to be equal or higher than the signal intensity of the tumour bed, a first criterion is considered to be met; compare a hemoglobin signal intensity of the spatial distribution of the hemoglobin with the signal intensity of the tumour bed or surroundings of the tumour bed; if the hemoglobin signal intensity at the margins is determined to be higher than the signal intensity of the tumour bed or the surroundings, a second criterion is considered to be met; if the collagen signal intensity or the hemoglobin signal intensity shows a direct continuity (or vascularity) to the tumour bed or at least part of a tumour with (or comprising) the tumour bed, a third criterion is considered to be met; and classify the margins based on at least one of the first criterion, the second criterion and the third criterion.

[0074] In one embodiment, the computing module 226 may be configured to: for no residual lesion observed from the real-time ultrasound images, if the spatial distribution of the lipids reflects disruption of a layer of the lipids and if all of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as positive margins; or alternatively, for no residual lesion observed from the real-time ultrasound images, if the spatial distribution of the lipids reflects disruption of the layer of the lipids and if none or only one or only two of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as negative margins.

[0075] The computing module 226 may further be configured to determine a presence of post-therapy changes at the margins, and if affirmative, deem the classified positive margins as a likelihood of false positive margins The prediction of the false positive margins may alternatively be performed or may additionally be affirmed by a relevant clinican.

[0076] In another embodiment, the computing module 226 may be configured to: for residual lesion observed from the real-time ultrasound images, if the spatial distribution of the lipids reflects disruption of a layer of the lipids and if only two or all of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as positive margins; or alternatively, for residual lesion observed from the realtime ultrasound images, if the spatial distribution of the lipids reflects disruption of the layer of the lipids and if none or only one of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as negative margins.

[0077] The computing module may further be configured to determine a presence of posttherapy changes at the margins and / or an extension of tumour cells into the margins, and if affirmative, deem the classified positive margins as a likelihood of false positive margins. The prediction of the false positive margins may alternatively be performed or may additionally be affirmed by a relevant clinican.

[0078] In yet another embodiment, the computing module is further configured to classify the margins as negative margins if the spatial distribution of the lipids reflects continuity or no disruption of a layer of the lipids.

[0079] In various embodiments, the apparatus 220 may be configured to operate in a plurality of modes.

[0080] Examples of the method 100 and the apparatus 220, along with a study to identify optoacoustic signatures / pattern of breast tissue post NACT, and to investigate its feasibility as biomarkers / prediction for margin assessment in the intra-operative setting, will be described in detail below.

[0081] The technology here lies in its combination of photoacoustic imaging and chromophore-based analysis to enable precise, real-time assessment of margins in post- NACT breast cancer specimens, thereby enhancing surgical outcomes and patient care. This may be applied across all photoacoustic imaging systems that may offer these chromophore information.MATERIALS AND METHODS

[0082] The study received approval from the local Institutional Review Board. From August 2020 to September 2021, women over 18 years old with histologically confirmed breast cancer who were referred to the Department of Surgery, National University Hospital, Singapore, for breast-conserving surgery (BCS) after neoadjuvant chemotherapy (NACT) were prospectively screened. Patients who underwent total mastectomy post- NACT or were deemed unsuitable for surgery (e.g. due to distant metastases) were excluded. After obtaining written informed consent, 21 women with histologically confirmed breast cancer referred for BCS post-NACTwere recruited. The data from the first four participants were allocated for training, while data from the remaining 17 were used for analysis.Equipment and Imaging Protocol

[0083] The excised breast tissue following lumpectomy was oriented using silk stitches by surgeons FIG. 10C shows a photograph of an exemplary specimen (excised breast tissue) 1011 with margins labelled using the silk stitches 1013, where a long stitch represents lateral (L) margin, a medium length stitch represents medial (M) margin, a short stitch represents superior (S) margin, and a loop stitch represents anterior (A) margin. Subsequently, the tissue underwent a saline rinse to eliminate surface blood. All specimens for this ex vivo investigation were collected within 10 minutes post-surgery from the operating theatre and promptly subjected to imaging to retain blood signals.

[0084] FIG. 10A shows a schematic view of a US-PA specimen imaging setup, according to an example. The setup included an optical parametric oscillator 1007 pumped by a Nd- YAG laser 1005, a handheld US-OT probe 1022 connected to the optical parametric oscillator 1007. The handheld US-OT probe 1022 was placed in heavy water tank 1009 on the sample (e g. 1011). For ultrasound-photoacoustic (US-PA) imaging, the multispectral optoacoustic tomography (MSOT) inVision 512-ECHO system (iThera Medical GmbH, Munich, Germany) 1026 was utilized, along with the handheld US-OT probe 1022 (which may be a specially designed handheld two-dimensional (2D) optoacoustic probe) capturing chromophores like lipids, collagen, and hemoglobin up to a 5 mm depth. As seen in FIG. 10B, this probe 1022 possessed an arc-shaped array of 256 detector elements with 125° angular coverage, arranged on a spherical surface (radius of about 40 mm) and operated at a central frequency of 5 MHz. Such system may be described in similar context to the ultrasound device 222 and the photoacoustic imaging module 224 of the apparatus 220 shown in FIG. 2.

[0085] To facilitate imaging, the specimen was positioned on a platform allowing for multidirectional movements — horizontal, vertical, and rotational. FIG. 10D shows a schematic view of the specimen 1011 illustrating the different margins based on the silk thread in an original positioning The top right inset of FIG 10D shows a schematic view of the specimen 1011 in a rotated position, while the bottom right inset of FIG. 10D shows a schematic view of the specimen 1011 in a flipped position (or flip side). For imaging the flip side, the specimen 1011 was manually flipped. Scanning commenced by positioning the probe 1022 on a computer-controlled stage (not shown in FIG. 10D), starting from the thickest part of the specimen 1011 and traversing it horizontally and vertically (placed on a silicone bed within an imaging chamber) through heavy water (D2O) (e.g. in the tank 1009) to ensure optimal acoustic coupling. Depending on size, the probe 1022 covered the entire specimen 1011 in increments of about 1 mm. Sections of ultrasound images, as seen in FIG. 10D, were acquired from medial (M) to lateral (L) margins.

[0086] Light for imaging was delivered via a fiber optic bundle integrated into the probe 1022, employing the wavelength-tunable optical parametric oscillator 1007 with selectable wavelengths ranging from 660 nm to 1300 nm, at a repetition rate of 10 Hz and per-pulse energy of 80 mJ (at about 730 nm). Various wavelengths (ranging from 700 nm to 1 100nm) were utilized to capture acoustic data from light-absorbing chromophores present in breast tissue (i.e. specimen 1011), including hemoglobin (Hb), blood oxyhemoglobin (HbO2), lipid, and collagen. The applied fluence remained below 20 mJ / cm2, adhering to the safety limit for near-infrared (NIR) nanosecond lasers at 10 Hz set by the American National Standards Institute. Real-time images generated by the back-projection algorithm were displayed during data acquisition.

[0087] Following US-PA imaging, specimens (e g. 1011) were immersed in formalin and transported to the laboratory for subsequent histopathological examination and for calculating diagnostic accuracy.Histopathological examination and juxtaposition with PA imaging

[0088] The plane of scan acquisition was conveyed to the pathologist, who sectioned the specimen accordingly to ensure accurate US-PA and histologic correlation. After postprocessing, the radiologist, PA scientist and pathologist gathered in a multi-disciplinary meeting to ascertain PA-pathology correlation by identifying the index / suspicious lesion and its relationship with adjacent margins. For patients that exhibited completed response, the clip that was in the tumour was identified as the center of lesion while for patients with nil / partial response, the residual tumour was identified as the region of interestImage reconstruction and analysis

[0089] FIG. 10E shows a normalized absorption spectra of endogenous chromophores in the excised breast specimen 1011 from the wavelength of 700 nm - HOO nm. Offline reconstruction of images acquired at each wavelength was conducted using default settings (back-projection algorithm, cut-off frequencies at 0.5 kHz to 6.5 MHz) on ViewMSOT 3.8. Spectral unmixing was performed using the default linear regression algorithm in ViewMSOT 3.8 to differentiate between chromophores, including Hb, HbO2, lipids, and collagen. Collagen and lipid signals were unmixed based on the entire spectral range (700 - 1100 nm), while Hb and HbO2 signals were calculated from a sub-range (700-850 nm) for enhanced accuracy in unmixing due to lower water absorptivity at these wavelengths. Color maps were assigned to each chromophore: lipids were represented in green, collagen in magenta, and total hemoglobin (HbT) in red. To optimize image processing andacquisition, US-PA images from the first participant in each category were evaluated by two unbhnded readers, YG (a breast radiologist with 6 years of experience in photoacoustic imaging) and GB (with 8 years of experience in clinical photoacoustic imaging).

[0090] Tissue samples were categorized into two groups for analysis: (1 ) cases indicating pathological complete response (pCR) with no residual US mass, and (2) cases with residual US mass suggesting incomplete / partial response. Cases showing ill-defined heterogeneous changes post-NACT with no definite three dimensions of a mass-like lesion on US were classified into group (I), where the original tumour bed was identified by the localization / tissue marker clip inserted prior to NACT. In group (2), the tumour bed was identified by the presence of a residual mass with definite 3D dimensions.

[0091] Optoacoustic distribution maps were generated for each specimen, and the results were scrutinized to pinpoint optoacoustic patterns / signatures specific to each group. Initially, for training purposes, US-PA images from the initial 2 participants in each group (totaling 4 participants) were collectively evaluated by a panel of three unblinded readers: YG, GB, and MO (a breast radiologist with 9 years of experience in clinical photoacoustic imaging). Following this, US-PA images for the remaining 17 participants were individually assessed by YG and GB, with any disparities resolved by MO, who was blinded to histopathologic findings.

[0092] In total, 102 margins were evaluated in excised specimens from the 17 study participants, with each specimen including 6 margins (anterior, posterior, superior, inferior, medial, and lateral). Presently, chromophore signals are subjected to visual and qualitative evaluation, though there are intentions to devise a quantitative assessment method in subsequent studies.Statistical Analysis

[0093] The results from margin assessments using ultrasound-photoacoustic (US-PA) imaging (e.g. carried out by the computing module 226 of FIG. 2) were compared to histopathologic findings in a categorical manner (i.e. positive / negative). Cross-tabulation was conducted for all assessed margins. The overall diagnostic accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of US-PA for margin diagnosis were calculated using SPSS software (Version 22; PASW Statistics, Chicago,Ill). Kappa statistics were determined to assess the overall agreement of US-PA with histopathology and to evaluate inter-observer variation in the assessment of (JS-PA images between YG and GB. Given the exploratory nature of this pilot proof-of-concept study, no sample size calculation was performed.RESULTSPatient characteristics[0094J Table 1 displays the baseline characteristics, histopathology of breast tumours, and the size of lumpectomy specimens from the 17 participants included in the analysis, with a mean age of 57.0 years ± 8.4.

[0095] TABLE 1: Baseline characteristics of patients (n = 17)Photoacoustic patterns observed across the NACT specimens[0096J The photoacoustic distribution maps of lipid, collagen and hemoglobin from the specimens were acquired up to a depth of 5 mm (data not shown).

[0097] The first PA pattern identified was the presence of intense and continuous lipid signals along the margins. A representative NACT specimen with partial / incomplete treatment response showing this pattern is presented in FIGS. 3A to 3F. FIG. 3A shows a clinical in-vivo ultrasound image depicting presence of an irregular hypoechoic mass (denoted by arrowheads in FIG. 3A) in a left breast of a patient (which was biopsied) proven to be a grade 3 invasive carcinoma (no special type). The patient subsequently underwent NACT. FIG. 3B shows an ex-vivo US image of the surgically excised specimen depicting a small residual hypoechoic mass (see arrows in FIG 3B) which is in keeping with partial / incomplete response. FIG. 3C shows a haematoxylin and eosin (H&E) stained microscopic image of specimen depicting residual tumour (see arrows in FIG. 3C) that is at least 3 mm away from all margins. The boxed area 301 corresponds to PA signals shown in FIGS. 3D to 3F.

[0098] FIG. 3D, FIG. 3E and FIG. 3F respectively show the corresponding total lipid, hemoglobin, and collagen distribution maps generated from PA imaging (on the left). On the right of each of FIG. 3D, FIG. 3E and FIG. 3F, the respective inverses are presented for better illustration. The presence of a continuous lipid signal (area enclosed by dotted line 303) in FIG. 3D is highly indicative of the presence of negative margins. The dotted line 303 is propagated to the hemoglobin, and collagen distribution maps of FIGS. 3E and 3F, respectively. The presence of some blood signals (see arrow) in FIG. 3E is normal and expected from post-surgical status / bleeding. The presence of high intensity blood and collagen signals (see * in FIGS. 3E and 3F) at the periphery of the excised tissue is due to orientation stitches.

[0099] The continuous lipid pattern (FIG. 3D), noted at the margins, is highly indicative of negative / clear margins with sensitivity, specificity, positive predictive value, or negative predictive value of 100%, 54.8%, 51.7%, 100% respectively. In areas of lipid discontinuity / disruption, optoacoustic distribution maps invariably identified collagen and hemoglobin signals. Areas with no lipid, collagen or hemoglobin signals were constantly related to imaging artefacts (e.g. air bubbles).

[0100] The optoacoustic distribution maps for collagen and hemoglobin at the margins, which resulted in discontinuity of lipid layer, demonstrated three patterns at the specimen margins when compared to the tumour bed. These patterns and the respective criteria are as described as follows.

[0101] (A) Intensity of collagen signal at margins cs tumour bed

[0102] If collagen signals are present at the margins, its signal intensity is directly compared to the signal intensity of the tumour bed. Overall findings are considered positive / suspicious if the collagen signals at the margins are equal or more intense as compared to the tumour bed, as these collagen signals are postulated to represent direct / indirect spread of disease towards the margins. If the collagen signals are less intense than the tumour bed, these findings are more likely to represent normal breast tissue on histology.

[0103] (B) Intensity of hemoglobin signal at margins

[0104] After inspection of collagen signals at the margins and comparing its intensity against the tumour bed, the vascularity of these regions were also evaluated with the use of hemoglobin signals. If the hemoglobin signals are more intense compared to the tumour bed or its surroundings, findings are deemed as positive / suspicious as tumour cells are known to demonstrate increased vascularity due to angiogenesis. Areas of increased hemoglobin signals with no corresponding collagen signals are often due to presence of hematoma / bruises which are normal / expected from surgeries.

[0105] (C) Connectivity of collagen and hemogoblin signal at margins to the tumour bed

[0106] After inspection of collagen and hemoglobin signals at the margins as described in (A) and (B), the authors observed another optoacoustic pattern which is connectivity of these signals towards the tumour bed. If the collagen / hemoglobin signals show a direct connection towards or direct continuity to the tumour bed / tumour, findings are deemed to be suspicious and worrisome for direct spread of disease.Overall interpretation[0107J The overall interpretation of margins depends on the permutation of the abovementioned three patterns (A) to (C). These patterns should not be read / interpreted in isolation as the factors are interconnected. For example, areas of abnormal collagen intensity at the margins may be worrisome for positive margins. However, in the absence of increased vascularity or connectivity towards the tumour bed, these areas of abnormal collagen could represent other collagenous breast conditions such as scarring / fibrosis which is an expected finding after NACT.[0108J Hence, the interpretation of margins requires permutation of the three optoacoustic patterns (A) to (C), as described above.

[0109] For Group 1 specimen (no residual lesion suggestive of pathological complete response (pCR)), the index tumour has demonstrated complete response to NACT. The threshold for considering positive margins is hence higher. For the Group (1) specimen, if the lipid layer is disrupted, a margin would only be positive on US-PA if all 3 criteria from (A) to (C) are met (see FIGS. 4A to 4F). Margins are classified as negative if 2 or less criteria from (A) to (C) are met.

[0110] FIGS. 4A to 4F illustrate positive margin patterns on US-PA images in the representative Group 1 specimen (no residual lesion on imaging post-NACT, suggestive of pCR. FIG. 4A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 3.5 cm irregular hypoechoic mass (see arrow heads in FIG. 4 A) in the left breast of a patient (which was biopsied) proven to be a grade 2 invasive carcinoma (no special type). The patient subsequently underwent NACT. FIG. 4B shows an ex-vivo US post-NACT image depicting no residual lesion upon surgical excision, highly suggestive of pCR. There is a faint echogenic clip noted (see arrowhead in FIG. 4B) in keeping with the tumour bed location. FIG. 4C shows a H&E stained microscopic image of specimen depicting areas of fibrosis on low power view (see arrowheads in FIG. 4C), but there was residual high-grade DCIS (ductal carcinoma in-situ) extending towards the posterior margins (dotted box 401). The DCIS was 1.5mm away from the posterior margins and was deemed as positive margins (< 2mm for DCIS).

[0111] FIG. 4D, FIG. 4E and FIG. 4F respectively show the corresponding lipid, total hemoglobin (Hbt) and collagen distribution maps generated from PA imaging (on the left).On the right of each of FIG. 4D, FIG. 4E and FIG. 4F, the respective inverses are presented for better illustration. There is focal disruption of continuous lipid signal (see dotted box 403) in FIG. 4D. The dotted box 403 is propagated to the hemoglobin distribution map of FIG. 4E. There is presence if mildly increased Hbt (see FIG 4E) and collagen (see FIG. 4F) signals at the area of lipid disruption. Of note, the collagen shows direct connectivity (see arrows in FIG. 4F) to the tumour bed / clip (see arrowhead in FIG. 4F). The presence of all 3 PA imaging criteria (A) to (C) were met and findings are suspicious for focal involvement of the posterior margins, concurring with histological findings.[0112J For a Group 2 specimen (with residual US mass suggestive of incomp lete / partial response), the threshold for positive margins are slightly lower (aqs compared to the Group 1 speciment) as these lesions are not complete responders which raise the possibility of residual cancer cells. For the Group (2) specimen, if the lipid layer is disrupted, a margin is positive if 2 or more criteria from (A) to (C) are met (see FIGS. 5A to 5F).

[0113] FIGS. 5A to 5F illustrate positive margin patterns on US-PA images in a representative NACT specimen showing residual lesion post NACT, suggestive of partial / incomplete response (Group 2 specimen). FIG. 5A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 1.5 cm irregular hypoechoic mass (see arrowheads in FIG. 5A) in the left breast of a patent (which was biopsied) proven to be a grade 3 invasive carcinoma (no special type) with associated ductal carcinoma in-situ. The patient subsequently underwent NACT. FIG. 5B shows an ex-vivo US post-NACT image depicting presence of residual lesion (see arrows in FIG. 5B) upon surgical excision, in keeping with partial / incomplete response. There is seemingly normal heterogeneous breast tissue between the residual tumour and the imaged inferior margin. FIG. 5C shows a H&E stained microscopic image of specimen depicting residual tumour (see arrows in FIG. 5C) with some response to presurgical therapy. The tumour bed extends up to all margins grossly with DCIS extending very close to the inferior margin (about 1 mm) (see dotted box 501). The tumour cells (both invasive and in-situ) also extend very close to (1 mm or less than 1 mm away from) several other margins (anterior, suprerior, medial and lateral) microscopically.

[0114] FIG. 5D, FIG. 5E and FIG. 5F respectively show the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging (on the left). Onthe right of each of FIG. 5D, FIG. 5E and FIG. 5F, the respective inverses are presented for better illustration. There is thinning / disruption of the lipid layer at the superior margin (see arrowheads in FIG. 5D). The area of disruption is noted to contain high intensities of hemoglobin (see arrowhead in FIG. 5E) and collagen (see arrowhead in FIG. 5F) with direct connectivity to the tumour bed (arrowsin FIGS. 5E and 5F). The presence of all 3 PA imaging criteria for positive margins were met and hence, PA imaging findings are highly indicative of positive margins.

[0115] Margins would be classified as negative if <2 criteria from (A) to (C) are met (see FIGS. 6A to 6F).

[0116] FIGS. 6A to 6F illustrate negative margin patterns on US-PA images in a representative NACT specimen showing residual lesion post NACT, suggestive of partial / incomplete response (Group 2 specimen). FIG. 6A shows a clinical in-vivo ultrasound pre-NACT image depicting the presence of a 2.0 cm irregular hypoechoic mass (see arrowheads in FIG. 6A) in the right breast of a patent (which was biopsied) proven to be a grade 3 invasive carcinoma (no special type). The patient subsequently underwent NACT. FIG. 6B shows an ex-vivo US post NACT image depicting presence of residual lesion (see arrows in FIG. 6B) upon surgical excision, in keeping with partial / incomplete response. FIG. 6C shows a H&E stained microscopic image of specimen depicting residual tumour (see arrows in FIG. 6C) that is away from the margins, with surrounding fibrosis (tumour bed) extending closer to the margins (boxed area 601).

[0117] FIG. 6D, FIG. 6E and FIG. 6F respectively show the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging (on the left). On the right of each of FIG. 6D, FIG. 6E and FIG. 6F, the respective inverses are presented for better illustration. There is thinning / disruption of the lipid layer (see arrowhead in FIG. 6D) noted along the lateral margins of the specimen The area of disruption is noted to contain collagen which is directly connected to the tumour bed (see arrowhead in FIG. 6F). However, this region of collagen signal intensity is not equal or higher than the tumour bed (see arrows in FIG. 6F) and this region does not show increased vascularity as compared to the tumour bed. Hence, only 1 imaging PA criteria was met (collagen connectivity to tumour bed) and findings are supportive of negative margins.

[0118] FIG. 7 shows a flowchart 708 depicting the summary of the overall findings (inset: decision tree on assessment of surgical magins based on photoacoustic patterns), as discussed above. This summary may be an example that is included in Step 108 of FIG. 1.Diagnostic accuracy of US-PA

[0119] In consensus with the acceptable margins set by Society of Surgical Oncology as having “no ink on tumour” for IDC (invasive ductal carcinoma) and a 2-mm negative margin for DCIS, margins involved are considered to be positive when there was < 2 mm of normal breast tissue seen between the tumour and the excised breast tissue margin of concern.

[0120] Of the 102 margins from 17 participants that were assessed, 2 margins were excluded due to heavy staining by patent blue, rendering the PA image uninterpretable. The remaining 100 margins were assessed according to the respective diagnostic criteria (Group 1 specimen: 54 margins, Group 2 specimen: 46 margins).

[0121] For Group (1) specimen, the diagnostic accuracy of US-PA showed a diagnostic accuracy of 98.1% (sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of 100%, 98.1%, 50.0% and 100%, respectively).

[0122] For Group (2) specimen, the diagnostic accuracy of US-PA showed a diagnostic accuracy of 78.3% (sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of 100%, 67.7%, 60.0% and 100%, respectively).

[0123] Overall, US-PA achieved a combined diagnostic accuracy of 89.0% (Groups 1 + 2 specimens) with sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of 100%, 86.9%, 59.4% and 100%, respectively.

[0124] Interpretation of the relationship of these 3 criteria with positivity / negativity of margins is summarized in Table 2

[0125] Table 2: Interpretation of the relationship of these 3 criteria with positivity / negativity of margins|0126| There was no interpretive discrepancy between the two readers (YG and GB) for the remaining 100 margins assessed, indicating excellent inter-observer agreement (interobserver variability = 1, kappa = 1). Based on histopathological examination, there was no evidence of laser-induced tissue damage to the excised specimen. The average scan time for each specimen in our study was approximately 20 minutes.False positives|0127| False positives were primarily due to post-therapy fibrotic changes and extremely close tumour extensions (<2 mm). There were total of 1 1 false positive margins in this study (1 in Group 1 specimen and 10 in Group 2 specimen). For Group 1 specimen where there was complete response, the false positive margin was secondary to the presence of posttherapy fibrotic changes at the margin (see FIGS. 8A to 8F). There were no residual cancer cells at the area of fibrosis on histopathology

[0128] FIGS. 8A to 8F illustrate false positive margin on US-PA images in a representativeNACT specimen showing no residual lesion post NACT on imaging, suggestive ofpathological complete response (pCR) (Group 1 specimen). FIG. 8A shows a clinical in- vivo ultrasound pre-NACT image depicting the presence of a 2.0 cm irregular hypoechoic mass (see arrowheads in FIG. 8A) in the left breast of a patient (which was biopsied) proven to be a grade 3 invasive carcinoma (no special type). The patient subsequently underwent NACT. FIG. 8B shows an ex-vivo US post-NACT image depicting no residual lesion upon surgical excision, highly suggestive of pCR. An echogenic clip (see arrowhead in FIG. 8B) is seen, representative of the tumour bed location. FIG. 8C shows a H&E stained microscopic image of specimen depicting no residual tumour, consistent with complete response to presurgical therapy. Residual tumour bed was identified near the lateral margin (boxed area 801).

[0129] FIG 8D, FIG 8E, and FIG 8F respectively show the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging (on the left). On the right of each of FIG. 8D, FIG. 8E and FIG. 8F, the respective inverses are presented for better illustration. There is thinning of the lipid layer along the lateral margin of the excised tissue, interspersed by dark linear bands (dotted box 803). The dotted box 303 is propagated to the hemoglobin, and collagen distribution maps of FIGS. 8E and 8F, respectively. The region of lipid thinning is occupied by increased signals of collagen and hemoglobin (see arrows in FIGS 8F and 8E). The signals of these chromophores show connectivity to the tumour bed (see arrowheads in FIGS. 8E and 8F) and their intensities are equal or raised as compared to the tumour bed. All 3 PA imaging criteria for positive margins were met and hence, PA imaging findings are worrisome for positive margins. However, there was no residual tumour cells noted on histology. This is an example of a false positive finding due to fibrotic changes post-NACT abutting the surgical margins.

[0130] For Group 2 specimen, 7 / 10 (70%) of the false positive cases were a result of extremely close extension of tumour cells to the margin. The distances range from 1 mm - 2 mm with an average of 1.3 mm (see FIGS. 9A to 9F). The other 3 / 10 (30%) of the false positives were similar to Group 1 specimen where there was presence of post-therapy fibrotic changes at the margin.

[0131] FIGS. 9A to 9F illustrate false positive margin on US-PA images in a representative NACT specimen showing residual lesion post NACT, suggestive of partial / incomplete response (Group 2 specimen) FIG. 9A shows a clinical in-vivo ultrasound pre-NACTimage depicting the presence of a 3.0 cm lobulated heterogeneous mass (see arrowheads in FIG. 9A) in the left breast of a patient (which was biopsied) proven to be a grade 2 invasive carcinoma (no special type). The patient subsequently underwent NACT. FIG. 9B shows an ex-vivo US post-NACT image depicting presence of residual mass (see arrows in FIG. 9B) upon surgical excision, in keeping with partial / incomplete response. FIG. 9C shows a H&E stained microscopic image of specimen depicting residual tumour (see arrows in FIG. 9C) that is close to but not involving the posterior (1 mm away) and inferior (2 mm away) margins (dotted box 901). The tumour also extends close to but not involving other margins (2 mm clearance from medial and anterior margins, not shown here).

[0132] FIG. 9D, FIG. 9E, and FIG. 9F respectively show the corresponding lipid, total hemoglobin and collagen distribution maps generated from PA imaging (on the left). On the right of each of FIG. 9D, FIG. 9E and FIG. 9F, the respective inverses are presented for better illustration. There is disruption of the lipid layer noted along the medial margins of the specimen (see arrowhead in dooted box 903). The dotted box 903 is propagated to the hemoglobin, and collagen distribution maps of FIGS. 9E and 9F, respectively The area of disruption is noted to contain collagen which is directly connected to the tumour bed. The collagen and hemoglobin signal intensities are much higher than the tumour bed. In this scenario, all 3 PA imaging criteria were met, and findings are worrisome for positive margins. However, there was at least 1 mm of margin clearance between the tumour and the margins on histology. This is hence an example of false positive margins due to extremely close distance between the tumour cells and the margin.DISCUSSION

[0133] In the present study, the feasibility of using ultrasound-guided photoacoustic tomography (US-PA) was investigated and its accuracy in assessing the margins of BCS specimens post NACT was determined. The results have shown that US-PA was able to provide clear distribution and visualisation of lipids, collagen and hemoglobin up to a depth of about 5 mm, with about 89.0% accuracy to histopathology.

[0134] In particular, the high NPV of 100% makes US-PA a potential intra-operative tool for assessment of margins in BCS. The assurance of a negative margin obtained reduces the risks of re-operations and locoregional recurrence risks. The positive predictive valueis however relatively low at 59.4% in this study. The majority (7 / 11) of these may be explained by extremely close tumour margins (< 2mm). From a clinical perspective, this is understandable and may result in no change in clinical management. This is because surgeons would err on the side of caution and obtain wider margins in these instances to reduce risks of locoregional recurrence. For the remaining 4 / 11, the false positive cases were secondary to post chemotherapy fibrotic changes at the margins which mimicked the features of positive margins on US-PA. As these entities share similar US-PA features qualitatively, future quantitative studies which measure the intensity of chromophores may be helpful in differentiation between true and false positive margins.

[0135] The unique US-PA patterns / signatures of margin assessment as mentioned above are first described here. Tn this study, these novel US-PA patterns / signatures have also been shown to correlate accurately with histopathology. In addition, these optical signatures also coincide with other studies which investigated the tumour microenvironment / collagen compartments of breast cancer. For example, collagen connectivity between tumour and the margins was described as a feature in a prior study where the presence of long collagen fibres / bands around tumours which serve as a path for potential tumour cell migration or immune cell response was described. It is believed that these US-PA features may not only be useful in assessment of tumour margins but are also helpful in future studies which aim to explore the tumour microenvironment of breast cancers.

[0136] The US-PA patterns / signatures were also easy to interpret. The inter-observer agreement was 1.0, showing high correlation between readers. Coupled with the clear optical contrast between different chromophores and the high sensitivity of US-PA to endogenous chromophores, US-PA is emerging as a promising tool due to easy discrimination. US-PA is also postulated to have a gentle learning curve as radiologists would be basing the primary investigation on the already familiar US. US-PA is label-free and does not rely on any exogeneous contrast agents which are often toxic. US-PA is safe as no histopathologic evidence of laser-induced damage to the scanned specimens were found.

[0137] Furthermore, US-PA imaging demonstrated efficiency with an average scan time of 20 minutes per specimen. This notably contrasts with the prolonged duration typically associated with histopathologic examination of tissue specimens and remains well belowthe recommended cold ischemia time of one hour. These findings underscore the potential integration of US-PA into the intraoperative clinical setting, enabling prompt decisions regarding the necessity for tissue re-excision to ensure complete tumour excision. Such integration has the potential to mitigate the requirement for repeat surgical procedures, thereby reducing the risks associated with patient morbidity.

[0138] It should be noted that this feasibility study is not without limitations. Firstly, the probe was only able to resolve lipid, collagen and blood signals up to a maximum depth of 5 mm (spatial resolution: 200 pm). These findings may be related to a lack of active circulation due to the ex-vivo nature of study. These findings may potentially affect the interpretation as one may not accurately depict the connect! on / collagen bands from the margins to the residual tumour in a large sample where the tumour is deep-seated. These findings may be improved by devising dedicated imaging probes for breast (e g. dimensions of probe) or by modifying light delivery methods (e.g. transillumination). Secondly, the sample size (n=17) for this study was considerably small, necessitating a larger cohort study to quantitatively corroborate our findings. Nevertheless, it is believed that this study remains one of the first with the largest number of margins assessed (100 margins).

[0139] US-PA provided clear visualization of tissue components, accurately correlating with histopathology. More specifically, the findings demonstrate that the US-PA probe provided comprehensive structural and functional insights into the margins of lumpectomy specimens, effectively delineating tissue chromophores. The high NPV of this method minimizes the risk of re-operations and locoregional recurrence. Although the PPV was lower, it did not impact clinical management as surgeons typically excise wider margins in such cases. The sensitivity of US-PA imaging to collagen, hemoglobin, and lipid holds significant promise for intraoperative assessment of tumour margins in post-NACT patients, offering a safe, rapid, and accurate approach. The strong agreement between US- PA analysis and histopathologic interpretation also suggests its potential utility as a tool for intraoperative margin assessment.

[0140] While the invention has been particularly shown and described with reference to specific embodiments, it should be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scopeof the invention as defined by the appended claims. The scope of the invention is thus indicated by the appended claims and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced.

Claims

CLAIMS1. A method of assessing margins of a post-Neoadjuvant Chemotherapy (NACT) breast cancer specimen, the method comprising: obtaining real-time ultrasound images of the post-NACT breast cancer specimen; simultaneously performing photoacoustic imaging on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen; fused with the real-time ultrasound images, analyzing the spatial distributions of the plurality of chromophores along the tissue boundary of the post-NACT breast cancer specimen to determine margins; and classifying the margins into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer specimen.

2. The method as claimed in claim 1, wherein the classified margins are positive margins if one of the two distinct situations is where cancer cells are present along at least part of the tissue boundary of the post-NACT breast cancer specimen, and wherein the classified margins are negative margins if the other of the two distinct situations is where cancer cells are absent along the tissue boundary of the post-NACT breast cancer specimen.

3. The method as claimed in claim 1 or 2, wherein the plurality of chromophores comprises lipids, hemoglobin and collagen.

4. The method as claimed in claim 3, wherein the plurality of chromophores further comprises elastin, myoglobin or melanin5. The method as claimed in claim 3 or 4, wherein if the spatial distribution of the lipids reflects disruption of a layer of the lipids, classifying the margins comprises:comparing a collagen signal intensity from the spatial distribution of the collagen with a signal intensity of a tumour bed of the post-NACT breast cancer specimen; if the collagen signal intensity at the margins is determined to be equal or higher than the signal intensity of the tumour bed, a first criterion is considered met; comparing a hemoglobin signal intensity of the spatial distribution of the hemoglobin with the signal intensity of the tumour bed or surroundings of the tumour bed; if the hemoglobin signal intensity at the margins is determined to be higher than the signal intensity of the tumour bed or the surroundings, a second criterion is considered met; if the collagen signal intensity or the hemoglobin signal intensity at the margins shows a direct continuity to the tumour bed or at least part of a tumour with the tumour bed, a third criterion is considered met; and classifying the margins based on at least one of the first criterion, the second criterion and the third criterion.

6. The method as claimed in claim 5, wherein classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion comprises: for no residual lesion observed from the real-time ultrasound images, if all of the first criterion, the second criterion and the third criterion are considered met, classifying the margins as positive margins; or for no residual lesion observed from the real-time ultrasound images, if none or only one or only two of the first criterion, the second criterion and the third criterion are considered met, classifying the margins as negative margins.

7. The method as claimed in claim 6, wherein classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion further comprises determining a presence of post-therapy changes at the margins, and if affirmative, the classified positive margins present a likelihood of false positive margins.

8. The method as claimed in claim 5, wherein classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion comprises:for residual lesion observed from the real-time ultrasound images, if only two or all of the first criterion, the second criterion and the third criterion are considered met, classifying the margins as positive margins; or for residual lesion observed from the real-time ultrasound images, i if none or only one of the first criterion, the second criterion and the third criterion are considered met, classifying the margins as negative margins.

9. The method as claimed in claim 8, wherein classifying the margins based on the at least one of the first criterion, the second criterion and the third criterion further comprises determining a presence of post-therapy changes at the margins and / or an extension of tumour cells into the margins, and if affirmative, the classified positive margins present a likelihood of false positive margins.

10. The method as claimed in claim 9, wherein the extension comprises a distance ranging from 1 mm to 2 mm.

11. The method as claimed in claim 3 or 4, wherein classifying the margins comprises if the spatial distribution of the lipids reflects no disruption of a layer of the lipids, classifying the margins as negative margins.

12. The method as claimed in any one of claims 1 to 11, wherein the margins comprise superior margins, anterior margins, inferior margins, medial margins, lateral margins or any combinations thereof.13 The method as claimed in any one of claims 1 to 12, wherein the spatial distributions of the plurality of chromophores are obtainable upto a depth of 5 mm of a specimen excised from the patient with the post-NACT breast cancer specimen during the breast-conserving surgery.

14. The method as claimed in any one of claims 1 to 12, wherein the spatial distributions of the plurality of chromophores are obtainable beyond a depth of 5 mm of aspecimen excised from the patient with the post-NACT breast cancer specimen during the breast-conserving surgery.

15. An apparatus for assessing margins of a post-Neoadjuvant Chemotherapy (NACT) breast cancer specimen, the apparatus comprising: an ultrasound device configured to capture real-time ultrasound images of the post- NACT breast cancer specimen; a photoacoustic imaging module configured to perform photoacoustic imaging on the post-NACT breast cancer specimen to map spatial distributions of a plurality of chromophores across a tissue boundary of the post-NACT breast cancer specimen, wherein the ultrasound device and the photoacoustic imaging module are simultaneously operable; and a computing module configured to: analyze, when fused with the real-time ultrasound images, the spatial distributions of the plurality of chromophores along the tissue boundary of the post- NACT breast cancer specimen to determine margins; and classify the margins into one out of two distinct situations such that the classified margins provide real-time actionable information to a surgeon during performance of a breast-conserving surgery on a patient with the post-NACT breast cancer specimen.

16. The apparatus as claimed in claim 15, wherein the photoacoustic imaging module comprises: a radio frequency source configured to deliver radio frequency pulses into the post- NACT breast cancer specimen; an ultrasonic transducer configured to detect ultrasonic waves generated by the post-NACT breast cancer specimen upon absorption of the radio frequency pulses and conversion into heat; and a data collector in communication with the ultrasonic transducer, wherein the data collector is configured to analyze the detected ultrasonic waves to produce images comprising the spatial distributions of the plurality of chromophores.

17. The apparatus as claimed in claim 15 or 16, wherein the plurality of chromophores comprises lipids, hemoglobin and collagen; and wherein if the spatial distribution of the lipids is determined to reflect disruption of a layer of the lipids, the computing module is configured to: compare a collagen signal intensity from the spatial distribution of the collagen with a signal intensity of a tumour bed of the post-NACT breast cancer specimen; if the collagen signal intensity at the margins is determined to be equal or higher than the signal intensity of the tumour bed, a first criterion is considered to be met; compare a hemoglobin signal intensity of the spatial distribution of the hemoglobin with the signal intensity of the tumour bed or surroundings of the tumour bed; if the hemoglobin signal intensity at the margins is determined to be higher than the signal intensity of the tumour bed or the surroundings, a second criterion is considered to be met; if the collagen signal intensity or the hemoglobin signal intensity at the margins shows a direct continuity to the tumour bed or at least part of a tumour with the tumour bed, a third criterion is considered to be met; and classify the margins based on at least one of the first criterion, the second criterion and the third criterion.

18. The apparatus as claimed in claim 17, wherein the computing module is configured to: for no residual lesion observed from the real-time ultrasound images, if all of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as positive margins; orfor no residual lesion observed from the real-time ultrasound images, if none or only one or only two of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as negative margins.

19. The apparatus as claimed in claim 18, wherein the computing module is further configured to determine a presence of post-therapy changes at the margins, and if affirmative, deem the classified positive margins as a likelihood of false positive margins.

20. The apparatus as claimed in claim 17, wherein the computing module is configured to: for residual lesion observed from the real-time ultrasound images, if only two or all of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as positive margins, or for residual lesion observed from the real-time ultrasound images, if none or only one of the first criterion, the second criterion and the third criterion are considered to be met, classify the margins as negative margins.

21. The apparatus as claimed in claim 20, wherein the computing module is further configured to determine a presence of post-therapy changes at the margins and / or an extension of tumour cells into the margins, and if affirmative, deem the classified positive margins as a likelihood of false positive margins.

22. The apparatus as claimed in claim 17, wherein the computing module is further configured to classify the margins as negative margins if the spatial distribution of the lipids reflects no disruption of a layer of the lipids.

Images