Intraobserver and interobserver variability in distinguishing between endocervical and endometrial adenocarcinoma on problematic cases of cervical curettings

Classification of cervical biopsy free-text diagnoses through linear-classifier based natural language processing

Routine cervical cancer screening has significantly decreased the incidence and mortality of cervical cancer. As selection of proper screening modalities depends on well-validated clinical decision algorithms, retrospective review correlating cytology and HPV test results with cervical biopsy diagnosis is essential for validating and revising these algorithms to changing technologies, demographics, and optimal clinical practices. However, manual categorization of the free-text biopsy diagnosis into discrete categories is extremely laborious due to the overwhelming number of specimens, which may lead to significant error and bias. Advances in machine learning and natural language processing (NLP), particularly over the last decade, have led to significant accomplishments and impressive performance in computer-based classification tasks. In this work, we apply an efficient version of an NLP framework, FastText™, to an annotated cervical biopsy dataset to create a supervised classifier that can assign accurate biopsy categories to free-text biopsy interpretations with high concordance to manually annotated data (>99.6%). We present cases where the machine-learning classifier disagrees with previous annotations and examine these discrepant cases after referee review by an expert pathologist. We also show that the classifier is robust on an untrained external dataset, achieving a concordance of 97.7%. In conclusion, we demonstrate a useful application of NLP to a real-world pathology classification task and highlight the benefits and limitations of this approach.

The variable clinicopathological categories and role of human papillomavirus in cervical adenocarcinoma: A hospital based nation-wide multi-center retrospective study across China

We investigated HPV in adenocarcinoma presenting and managed as cervical adenocarcinoma (CADC) at seven major representative regional cancer centres across China. From 1,051 CADC cases diagnosed locally in 2005-2010, 881 had available paraffin embedded tissue. Initial review excluded 154 cases as other diagnoses or inappropriate specimens. In 718 eligible cases consensus panel pathology diagnosis was made using an algorithm incorporating p16 and progesterone receptor immunohistochemistry (IHC). Classification of cervical adenocarcinoma categories was subject to substantial pathological disagreement. High-risk human papillomaviruses (HR-HPV) DNA was studied by the sensitive SPF10 PCR-DEIA-LiPA25 version 1 for L1 genes and type-specific HR-HPV E6/7 gene PCR's. HR-HPV prevalence in whole tissue samples in eligible tested CADC was 74.5%: 100.0% in neuro-endocrine carcinoma (NEC), 82.2% in classical cervical adenocarcinoma (ADC-CX), 40.0% in adenocarcinoma-not otherwise specified (ADC-NOS) and 33.3% in endometrioid adenocarcinoma (ADC-ENDO). Higher mean age at diagnosis correlated with histological categories showing low HPV prevalence (Linear regression: β= -13.794, p < 0.001). HPV-16 and 18 were associated with early development of CADC and a lower mean age correlated with carcinogenic risk of associated HPV (β = -0.1829, p < 0.001). HPV-16 or HPV-18 was found in 88.2% of all HPV positive cases including multiple-infections. HPV-18 was the commonest HPV type in NEC (58.3%), ASC (40.2%) and ADC-CX (40.9%). The proportion of HPV-unrelated CADC and in different final histological categories varied geographically and by age. Although HPV negativity was predominantly associated with special categories of CADC, some HPV-negative usual adenocarcinomas indistinguishable by adjudicated microscopic diagnosis from ADC-CX were found and varied in frequency across China.

Reinventing diagnostics for personalized therapy in oncology

Human cancers are still diagnosed and classified using the light microscope. The criteria are based upon morphologic observations by pathologists and tend to be subject to interobserver variation. In preoperative biopsies of non-small cell lung cancers, the diagnostic concordance, even amongst experienced pulmonary pathologists, is no better than a coin-toss. Only 25% of cancer patients, on average, benefit from therapy as most therapies do not account for individual factors that influence response or outcome. Unsuccessful first line therapy costs Canada CAN$1.2 billion for the top 14 cancer types, and this extrapolates to $90 billion globally. The availability of accurate drug selection for personalized therapy could better allocate these precious resources to the right therapies. This wasteful situation is beginning to change with the completion of the human genome sequencing project and with the increasing availability of targeted therapies. Both factors are giving rise to attempts to correlate tumor characteristics and response to specific adjuvant and neoadjuvant therapies. Static cancer classification and grading systems need to be replaced by functional classification systems that not only account for intra- and inter- tumor heterogeneity, but which also allow for the selection of the correct chemotherapeutic compounds for the individual patient. In this review, the examples of lung and breast cancer are used to illustrate the issues to be addressed in the coming years, as well as the emerging technologies that have great promise in enabling personalized therapy.

Distinguishing between primary endocervical and endometrial adenocarcinomas: is a 2-marker (Vim/CEA) panel enough?

Gynecological pathologists are used to operating many panels of various markers in combination for the diagnostic distinction between primary endocervical and endometrial adenocarcinomas. The conventional 3-marker (ER/Vim/CEA) panel is the most promising tool. In this study, our aim is to investigate whether a 2-marker panel is enough to distinguish between these two gynecologic malignancies. Additionally, we wish to determine which one is the most favorable among eight panels tested, including six 2-marker (ER/CEA, PR/CEA, Vim/CEA, ER/p16(INK4a), PR/p16(INK4a), Vim/p16(INK4a)) and two 3-marker (ER/Vim/CEA, ER/Vim/p16(INK)) panels. A tissue microarray was constructed using paraffin-embedded, formalin-fixed tissues from 35 hysterectomy specimens, including 14 primary endocervical adenocarcinomas and 21 primary endometrial adenocarcinomas. Utilizing the avidin-biotin complex (ABC) method, tissue array sections were immunostained with five commercially available antibodies (ER, Vim, CEA, PR, and p16(INK4a)) to evaluate their individual frequencies of expression. We found that all eight aforementioned panels showed an encouraging range of overall accuracy (69.2% to 78.3%). However, one panel of 2-markers (Vim, CEA) exhibited the most efficiency (78.3%) in the diagnostic distinction between primary endocervical and endometrial adenocarcinomas. Based on the analyzed data, we conclude that the 2-marker (Vim/CEA) panel seems adequate to be an appropriate, convenient, and efficient means to distinguish between primary endocervical and endometrial adenocarcinomas. Even though there were a limited number of cases, this study still provides valuable references to help avoid wasting resources and unnecessary marker testing.