Friday, April 18, 2025
4/18/2025
Advancing Thyroid Cancer Diagnostics with AI-enhanced Multimodal Optical Histopathology - PROJECT SUMMARY/ ABSTRACT The increasing incidence of thyroid cancer worldwide is partly fueled by enhanced detection of smaller nodules through imaging techniques like ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI), leading to overdiagnosis of low-risk cancers and benign thyroid conditions. Papillary thyroid carcinoma, the most prevalent type of thyroid cancer, presents lymph node metastasis (LNM) in 60 to 70% of cases, increasing the risk of local recurrence, distant metastasis, and mortality. Fine needle aspiration (FNA), the primary diagnostic method, yields indeterminate results for 10 to 20% of nodules, of which 20 to 30% of nodules may be malignant. Such uncertainty results in repeat biopsies, costly molecular tests, or unnecessary surgeries, causing increased healthcare costs, reduced quality of life, and emotional distress. Core needle biopsies (CNB) are suggested for indeterminate FNA findings. However, CNB is also prone to sampling errors, not suitable near critical structures, and both FNA and CNB are time-consuming procedures and can alter tissue morphology. Given the prevalence of false positives, complications arising from indeterminate diagnoses, prolonged analysis periods, and sampling challenges in thyroid cancer detection, there is a pressing demand for a more effective diagnostic solution. Our proposal introduces an AI-augmented, multimodal, label-free nonlinear optical microscopy system that incorporates coherent anti-Stokes Raman scattering microscopy (CARS), second harmonic generation microscopy (SHG), and two-photon autofluorescence microscopy (TPAF). This innovative system aims for rapid and precise diagnosis of thyroid cancer and LNM. By eliminating the need for external dyes, the label-free technique streamlines the diagnostic process and minimizes complications. This combination of imaging modalities with AI augmentation offers detailed and complimentary subcellular morphological and biochemical information, enhancing our understanding of biochemical composition of thyroid nodules and lymph nodes, facilitating identification of malignancy markers more accurately, and significantly improving diagnostic accuracy. In addition, we plan to develop a compact, portable, AI integrated label-free microendoscope for clinical translation, designed to fit within a core biopsy needle, for differentiating between cancerous and normal tissues and detecting LNM without tissue excision. Validated through studies on ex vivo human tissues and in vivo porcine models, this device promises to transform thyroid cancer diagnosis by serving as a virtual histopathology tool, enabling real-time, label-free imaging and addressing the drawbacks of FNA and CNB, such as overdiagnoses, indeterminate diagnoses, and false positives, without tissue excision or harm to critical adjacent structures. Upon project completion, we anticipate delivering an AI-enhanced, label-free benchtop system for testing in outpatient clinics, capable of imaging FNA samples for rapid and precise diagnosis within minutes, and a comparable label-free microendoscope as a potential optical histopathology device to facilitate near real-time cancer diagnosis without tissue excision.
HHS’ Tracking Accountability in Government Grants System (TAGGS) website provides access to detailed descriptions of grants, loans, aggregated direct payments and other types of financial assistance awarded by the HHS Operating Divisions (OPDIVs) and the Office of the Secretary Staff Divisions (STAFFDIVs).