scholarly journals Collaborative workflow between pathologists and deep learning for evaluation of tumor cellularity in lung adenocarcinoma

2022 ◽  
Author(s):  
Taro Sakamoto ◽  
Tomoi Furukawa ◽  
Hoa H.N. Pham ◽  
Kishio Kuroda ◽  
Kazuhiro Tabata ◽  
...  
Keyword(s):  
Ground Truth ◽  
Quantitative Model ◽  
Cell Counting ◽  
Tumor Segmentation ◽  
Molecular Testing ◽  
Routine Practice ◽  
Thyroid Transcription Factor 1 ◽  
Thyroid Transcription Factor ◽  
Adjusted Score ◽  
Adjusted Scores

Owing to the high demand for molecular testing, the reporting of tumor cellularity in cancer samples has become a mandatory task for pathologists. However, the pathological estimation of tumor cellularity is often inaccurate. We developed a collaborative workflow between pathologists and artificial intelligence (AI) models to evaluate tumor cellularity in lung cancer samples and prospectively applied it to routine practice. We also developed a quantitative model that we validated and tested on retrospectively analyzed cases and ran the model prospectively in a collaborative workflow where pathologists could access the AI results and apply adjustments (Adjusted-Score). The Adjusted-Scores were validated by comparing them with the ground truth established by manual annotation of hematoxylin-eosin slides with reference to immunostains with thyroid transcription factor-1 and napsin A. For training, validation, retrospective testing, and prospective application of the model, we used 40, 10, 50, and 151 whole slide images, respectively. The sensitivity and specificity of tumor segmentation were 97% and 87%, and the accuracy of nuclei recognition was 99%. Pathologists altered the initial scores in 87% of the cases after referring to the AI results and found that the scores became more precise after collaborating with AI. For validation of Adjusted-Score, we found the Adjusted-Score was significantly closer to the ground truth than non-AI-aided estimates (p<0.05). Thus, an AI-based model was successfully implemented into the routine practice of pathological investigations. The proposed model for tumor cell counting efficiently supported the pathologists to improve the prediction of tumor cellularity for genetic tests.

Immunohistochemistry Profile Predicts EGFR Mutation Status in Lung Adenocarcinoma

2020 ◽  
Vol 28 (5) ◽  
pp. 502-506
Author(s):  
Wencheng Li ◽  
Angela G. Niehaus ◽  
Stacey S. O’Neill
Keyword(s):  
Lung Adenocarcinoma ◽  
Egfr Mutation ◽  
Growth Factor Receptor ◽  
Egfr Mutations ◽  
Molecular Testing ◽  
Mutation Status ◽  
Thyroid Transcription Factor 1 ◽  
Thyroid Transcription Factor ◽  
Napsin A ◽  
Therapeutic Decisions

Significant advances in targeted therapy have been made in recent years for patients with lung adenocarcinoma. These targeted therapies have made molecular testing of paramount importance to drive therapeutic decisions. Material for testing is often limited, particularly in cytology specimens and small core biopsies. A reliable screening tool is invaluable in triaging limited tissue and selection for epidermal growth factor receptor ( EGFR) mutation testing. We hypothesized that the immunohistochemistry (IHC) profile of lung adenocarcinoma predicts EGFR mutation status. In this retrospective study, we evaluated the thyroid transcription factor-1 (TTF-1)/napsin A IHC profile and EGFR mutation status in 339 lung adenocarcinomas at our academic institution. In our cohort, we found that 92.3% of cases were positive for TTF-1 and/or napsin A by IHC with an EGFR positivity rate of 17.3%. Importantly, 7.7% of the cases were dual TTF-1/napsin A negative, and none of these cases contained EGFR mutations. This finding supports the use of TTF-1 and napsin A IHC to identify cases where EGFR mutation status will be negative, thus preserving limited tissue for other ancillary testing.

scholarly journals Differential Diagnosis of Epithelioid Malignant Mesothelioma With Lung and Breast Pleural Metastasis: A Systematic Review Compared With a Standardized Panel of Antibodies—A New Proposal That May Influence Pathologic Practice

2019 ◽  
Vol 144 (4) ◽  
pp. 446-456 ◽  
Author(s):  
Nolwenn Le Stang ◽  
Louise Burke ◽  
Gaetane Blaizot ◽  
Allen R. Gibbs ◽  
Pierre Lebailly ◽  
...  
Keyword(s):  
Systematic Review ◽  
Differential Diagnosis ◽  
Malignant Mesothelioma ◽  
Estrogen Receptor Α ◽  
Routine Practice ◽  
Breast Cancers ◽  
Thyroid Transcription Factor 1 ◽  
Thyroid Transcription Factor ◽  
Pleural Metastasis ◽  
Transcription Factor 1

Context.— Pleural mesothelioma is a rare cancer with an often-challenging diagnosis because of its potential to be a great mimicker of many other tumors. Among them, primary lung and breast cancers are the 2 main causes of pleural metastasis. The development and application of targeted therapeutic agents have made it even more important to achieve an accurate diagnosis. In this setting, international guidelines have recommended the use of 2 positive and 2 negative immunohistochemical biomarkers. Objectives.— To define the most highly specific and sensitive minimum set of antibodies for routine practice to use for the separation of epithelioid malignant mesothelioma from lung and breast metastasis and to determine the most relevant expression cutoff. Design.— To provide information at different levels of expression of 16 mesothelial and epithelial biomarkers, we performed a systematic review of articles published between 1979 and 2017, and we compared those data to results from the Mesothelioma Telepathology Network (MESOPATH) of the standardized panel used in routine practice database since 1998. Results.— Our results indicate that the following panel of markers—calretinin (poly)/thyroid transcription factor 1 (TTF-1; clone 8G7G3/1) and calretinin (poly)/estrogen receptor-α (ER-α; clone EP1)—should be recommended; ultimately, based on the MESOPATH database, we highlight their relevance which are the most sensitive and specific panel useful to the differential diagnosis at 10% cutoff. Conclusions.— Highlighted by their relevance in the large cohort reported, we recommend 2 useful panels to the differential diagnosis at 10% cutoff.

scholarly journals IMG-22. A DEEP LEARNING MODEL FOR AUTOMATIC POSTERIOR FOSSA PEDIATRIC BRAIN TUMOR SEGMENTATION: A MULTI-INSTITUTIONAL STUDY

2020 ◽  
Vol 22 (Supplement_3) ◽  
pp. iii359-iii359
Author(s):  
Lydia Tam ◽  
Edward Lee ◽  
Michelle Han ◽  
Jason Wright ◽  
Leo Chen ◽  
...  
Keyword(s):  
Deep Learning ◽  
Posterior Fossa ◽  
Ground Truth ◽  
Learning Model ◽  
Tumor Segmentation ◽  
Dice Similarity Coefficient ◽  
Study Cohort ◽  
Navigation Systems ◽  
Tumor Types ◽  
Deep Learning Model

Abstract BACKGROUND Brain tumors are the most common solid malignancies in childhood, many of which develop in the posterior fossa (PF). Manual tumor measurements are frequently required to optimize registration into surgical navigation systems or for surveillance of nonresectable tumors after therapy. With recent advances in artificial intelligence (AI), automated MRI-based tumor segmentation is now feasible without requiring manual measurements. Our goal was to create a deep learning model for automated PF tumor segmentation that can register into navigation systems and provide volume output. METHODS 720 pre-surgical MRI scans from five pediatric centers were divided into training, validation, and testing datasets. The study cohort comprised of four PF tumor types: medulloblastoma, diffuse midline glioma, ependymoma, and brainstem or cerebellar pilocytic astrocytoma. Manual segmentation of the tumors by an attending neuroradiologist served as “ground truth” labels for model training and evaluation. We used 2D Unet, an encoder-decoder convolutional neural network architecture, with a pre-trained ResNet50 encoder. We assessed ventricle segmentation accuracy on a held-out test set using Dice similarity coefficient (0–1) and compared ventricular volume calculation between manual and model-derived segmentations using linear regression. RESULTS Compared to the ground truth expert human segmentation, overall Dice score for model performance accuracy was 0.83 for automatic delineation of the 4 tumor types. CONCLUSIONS In this multi-institutional study, we present a deep learning algorithm that automatically delineates PF tumors and outputs volumetric information. Our results demonstrate applied AI that is clinically applicable, potentially augmenting radiologists, neuro-oncologists, and neurosurgeons for tumor evaluation, surveillance, and surgical planning.

scholarly journals Identification of Thyroid Transcription Factor-1 in C Cells and Parathyroid Cells

Endocrinology ◽  
1998 ◽  
Vol 139 (6) ◽  
pp. 3014-3017 ◽  
Author(s):  
Koichi Suzuki ◽  
Yoshihiko Kobayashi ◽  
Ryohei Katoh ◽  
Leonard D. Kohn ◽  
Akira Kawaoi
Keyword(s):  
Transcription Factor ◽  
C Cells ◽  
Thyroid Transcription Factor 1 ◽  
Thyroid Transcription Factor ◽  
Transcription Factor 1

Thyroid transcription factor-1 expression in colorectal adenocarcinomas

2011 ◽  
Vol 207 (11) ◽  
pp. 686-690 ◽  
Author(s):  
Matthias Dettmer ◽  
Tae Eun Kim ◽  
Chan Kwon Jung ◽  
Eun Sun Jung ◽  
Kyo Young Lee ◽  
...  
Keyword(s):  
Transcription Factor ◽  
Thyroid Transcription Factor 1 ◽  
Thyroid Transcription Factor ◽  
Transcription Factor 1 ◽  
Colorectal Adenocarcinomas

Thyroid Transcription Factor 1 Immunohistochemistry as an Intraoperative Diagnostic Tool at Frozen Section for Distinction Between Primary and Secondary Lung Tumors

2007 ◽  
Vol 131 (4) ◽  
pp. 582-587
Author(s):  
David N. Butcher ◽  
Peter Goldstraw ◽  
George Ladas ◽  
Michael E. Dusmet ◽  
Mary N. Sheppard ◽  
...  
Keyword(s):  
Transcription Factor ◽  
Frozen Section ◽  
Operative Procedure ◽  
Tumor Type ◽  
Diagnostic Confidence ◽  
Thyroid Transcription Factor 1 ◽  
Specificity And Sensitivity ◽  
Thyroid Transcription Factor ◽  
Transcription Factor 1 ◽  
Pulmonary Tumors

Abstract Context.—Intraoperative distinction between primary and metastatic carcinomas in the lung at frozen section remains problematic. Objective.—To assess the value and practicality of immunohistochemistry for thyroid transcription factor 1 at the time of intraoperative frozen section. Design.—Thirty-three patients presented with either a solitary pulmonary mass or 2 pulmonary masses and a history of carcinoma in a different organ. In addition to routine frozen section for assessment of tumor type, we looked for expression of thyroid transcription factor 1, using the EnVision system with abridged methodology. Results.—Ten cases were positive for thyroid transcription factor 1, which was confirmed on subsequent paraffin sections. Nine of these were confirmed as primary pulmonary adenocarcinomas, but 1 case proved to be a rare false-positive metastatic colonic carcinoma. Twenty-three cases were negative on frozen section and reported as favoring metastatic disease. In all cases, additional immunohistochemical data increased diagnostic confidence, but particularly in cases of positive primary pulmonary tumors and in cases with disease metastatic from sites other than the large bowel. The average time in addition to that of the basic frozen section was 24 minutes per test with a cost of £32 (US$57). Conclusions.—Frozen section immunohistochemistry for thyroid transcription factor 1 shows specificity and sensitivity similar to those seen for formalin-fixed tissues and is feasible within the time frame of a thoracotomy. Diagnostic confidence is increased, especially with positive primary pulmonary tumors. However, its practice should be properly planned within an operative procedure as liberal usage will likely have significant staff and cost implications.

Sign in / Sign up

Export Citation Format

Share Document