Current status of machine learning in thyroid cytopathology.

The implementation of Digital Pathology has allowed the development of computational Pathology. Digital image-based applications that have received FDA Breakthrough Device Designation have been primarily focused on tissue specimens. The development of Artificial Intelligence-assisted algorithms using Cytology digital images has been much more limited due to technical challenges and a lack of optimized scanners for Cytology specimens. Despite the challenges in scanning whole slide images of cytology specimens, there have been many studies evaluating CP to create decision-support tools in Cytopathology. Among different Cytology specimens, thyroid fine needle aspiration biopsy (FNAB) specimens have one of the greatest potentials to benefit from machine learning algorithms (MLA) derived from digital images. Several authors have evaluated different machine learning algorithms focused on thyroid cytology in the past few years. The results are promising. The algorithms have mostly shown increased accuracy in the diagnosis and classification of thyroid cytology specimens. They have brought new insights and demonstrated the potential for improving future cytopathology workflow efficiency and accuracy. However, many issues still need to be addressed to further build on and improve current MLA models and their applications. To optimally train and validate MLA for thyroid cytology specimens, larger datasets obtained from multiple institutions are needed. MLAs hold great potential in improving thyroid cancer diagnostic speed and accuracy that will lead to improvements in patient management.

Primary cutaneous SMARCB1-deficient carcinoma.

BACKGROUND: SMARCB1-deficient malignancies can arise in various sites. We describe a novel primary SMARCB1-deficient carcinoma of skin (SDCS) and characterize SMARCB1 mutations in non-melanoma skin cancers (NMSC). METHODS: Cases underwent immunophenotyping and targeted exome sequencing (MSK-IMPACT) assay interrogating somatic mutations in 468 cancer-related genes. The MSK-IMPACT database from 2014 to 2020 encompassing 55, 000 cases was searched for NMSC with SMARCB1 mutations. RESULTS: SDCS arose on the scalp of an 18-year-old woman showing homozygous SMARCB1 deletion with a LATS2 G963E variant. Another case arose on the temple of a 76-year-old man harboring a SMARCB1 W206* mutation associated with loss of heterozygosity (LOH), 59 concurrent mutations, and a UV mutation signature (UV-MS). Both tumors exhibited INI1 loss, positive CK5/6, p40, p63, and claudin-4 with negative CD34. Of 378 NMSC cases, including 370 carcinomas, 7 SMARCB1-mutated tumors were identified: 3 squamous cell, 3 Merkel cell, and one basal cell carcinoma. Six showed UV-MS. Five INI1-interrogated cases retained protein expression suggesting they were SMARCB1-proficient. CONCLUSIONS: SDCS can be clinically aggressive, harbor SMARCB1 homozygous deletions or truncating SMARCB1 mutations associated with LOH, and can occur with or without UV-MS. Overall, SMARCB1 mutations in NMSC are rare with most being of undetermined significance and associated with retained INI1 and UV-MS.

SMARCB1-deficient carcinomas of the head and neck region: a cytopathologic characterization.

INTRODUCTION: SMARCB1 encodes for a component of the SWI/SNF complex and is widely implicated in carcinogenesis. In the head and neck, SMARCB1-deficient carcinomas typically arise in the sinonasal tract but can be found at other sites. EZH2 inhibitors have emerged as potential targeted therapy against SWI/SNF-deficient tumors. We sought to characterize the cytomorphology of head and neck carcinomas with SMARCB1 deficiencies to identify potential candidates for targeted therapy. MATERIALS AND METHODS: Head and neck carcinomas with SMARCB1 mutations were retrospectively identified and confirmed to be SMARCB1-deficient by both molecular (fluorescent in-situ hybridization or next generation sequencing) and immunohistochemical means. Cases with positive cytology were reviewed and their cytologic features cataloged. RESULTS: A total of 19 specimens from 13 patients were reviewed, including 8 specimens from 7 sinonasal carcinomas, 4 specimens from 3 thyroid carcinomas, 3 specimens from 2 skin carcinomas, and 4 specimens from 1 carcinoma of unknown primary origin. High-grade features were common, including mitoses (11 of 19) necrosis (13 of 19) and multinucleation (16 of 19). Tumors showed either dense cytoplasm with distinct cell borders (10 of 19) or delicate cytoplasm with indistinct cell borders (9 of 19). Most tumors showed no distinct epithelial differentiation (12 of 19), while some (7 of 19) showed glandular or signet ring features. A minor cohort demonstrated rhabdoid cells (4 of 19). CONCLUSIONS: Head and neck carcinomas with SMARCB1 deficiencies have a wide array of morphologies and tend to demonstrate high-grade features. Only a minor cohort demonstrate rhabdoid-type cells. Evaluation of SMARCB1 deficiency for potential targeted therapy should not be limited to tumors with rhabdoid morphology.

Vasitis nodosa and related lesions: a modern immunohistochemical staining profile with special emphasis on novel diagnostic dilemmas.

Vasitis nodosa is a benign proliferation of vas deferens epithelium, thought to be a response to trauma or obstruction, usually vasectomy. Although histologic features mimic malignancy, diagnosis is usually straightforward due to the clinical context. We analyzed 21 specimens with vasitis or epididymitis nodosa with antibodies to PAX8, CD10, p63, α-methyl-acyl-coA-racemase (AMACR), GATA3, prostein, NKX3.1, and prostate-specific antigen (PSA). Two diagnostically problematic cases included (1) florid bladder muscle involvement after prostatectomy and (2) involvement of the ampulla and ejaculatory duct in a radical prostatectomy specimen. Vasitis nodosa was excluded in 3 additional histologic mimics (2 post-treatment prostate cancers and 1 bladder cancer). PAX8 yielded consistent positive (100%) nuclear staining in the proliferative glands of vasitis nodosa, often stronger and more uniform than native vas deferens. CD10 labeling was common but also labeled secretions and other structures. Labeling for p63 was often basally located in glands with a multilayered appearance, but often markedly attenuated or lacking in the proliferative glands compared to native epithelium. AMACR positivity was variable but often present (19/21). PSA, prostein, and NKX3.1 were consistently negative. Rare problematic cases of vasitis nodosa include "invasion" of the ejaculatory duct at the prostate and involvement of bladder muscle after prostatectomy. The proliferative vasitis nodosa glands often have a prostate cancer-like staining pattern with variable AMACR positivity and negative or patchy p63. However, reliable positivity for PAX8, patchy GATA3, and negative staining for PSA, NKX3.1, and prostein aid in distinguishing from prostate cancer and tubular variants of bladder cancer.