The difficulty with measuring the largest melanoma tumour diameter in sentinel lymph nodes.

Identification of sentinel node (SN) metastases can set the adjuvant systemic therapy indication for stage III melanoma patients. For stage IIIA patients, a 1.0 mm threshold for the largest SN tumour diameter is used. Therefore, uniform reproducible measurement of its size is crucial. At present, the number of deposits or their microanatomical sites are not part of the inclusion criteria for adjuvant treatment. The goal of the current study was to show examples of the difficulty of measuring SN melanoma tumour diameter and teach how it should be measured. Histopathological slides of SN-positive melanoma patients were retrieved using the Dutch Pathology Registry (PALGA). Fourteen samples with the largest SN metastasis around 1.0 mm were uploaded via tele-pathology and digitally measured by 12 pathologists to reflect current practice of measurements in challenging cases. Recommendations as educational examples were provided. Microanatomical location of melanoma metastases was 1 subcapsular, 2 parenchymal and 11 combined. The smallest and largest difference in measurements were 0.24 mm and 4.81 mm, respectively. 11/14 cases (78.6%) showed no agreement regarding the 1.0 mm cut-off. The median discrepancy for cases ≤5 deposits was 0.5 mm (range 0.24-0.60, n=3) and 2.51 mm (range 0.71-4.81, n=11) for cases with ≥6 deposits. Disconcordance in measuring SN tumour burden is correlated with the number of deposits. Awareness of this discordance in challenging cases, for example, cases with multiple small deposits, is important for clinical management. Illustrating cases to reduce differences in size measurement are provided.

Rocky road to digital diagnostics: implementation issues and exhilarating experiences.

Since 2007, we have gradually been building up infrastructure for digital pathology, starting with a whole slide scanner park to build up a digital archive to streamline doing multidisciplinary meetings, student teaching and research, culminating in a full digital diagnostic workflow where we are currently integrating artificial intelligence algorithms. In this paper, we highlight the different steps in this process towards digital diagnostics, which was at times a rocky road with definitely issues in implementation, but eventually an exciting new way to practice pathology in a more modern and efficient way where patient safety has clearly gone up.

Diagnostic Assessment of Deep Learning Algorithms for Detection of Lymph Node Metastases in Women With Breast Cancer.

Importance: Application of deep learning algorithms to whole-slide pathology images can potentially improve diagnostic accuracy and efficiency. Objective: Assess the performance of automated deep learning algorithms at detecting metastases in hematoxylin and eosin-stained tissue sections of lymph nodes of women with breast cancer and compare it with pathologists' diagnoses in a diagnostic setting. Design, Setting, and Participants: Researcher challenge competition (CAMELYON16) to develop automated solutions for detecting lymph node metastases (November 2015-November 2016). A training data set of whole-slide images from 2 centers in the Netherlands with (n = 110) and without (n = 160) nodal metastases verified by immunohistochemical staining were provided to challenge participants to build algorithms. Algorithm performance was evaluated in an independent test set of 129 whole-slide images (49 with and 80 without metastases). The same test set of corresponding glass slides was also evaluated by a panel of 11 pathologists with time constraint (WTC) from the Netherlands to ascertain likelihood of nodal metastases for each slide in a flexible 2-hour session, simulating routine pathology workflow, and by 1 pathologist without time constraint (WOTC). Exposures: Deep learning algorithms submitted as part of a challenge competition or pathologist interpretation. Main Outcomes and Measures: The presence of specific metastatic foci and the absence vs presence of lymph node metastasis in a slide or image using receiver operating characteristic curve analysis. The 11 pathologists participating in the simulation exercise rated their diagnostic confidence as definitely normal, probably normal, equivocal, probably tumor, or definitely tumor. Results: The area under the receiver operating characteristic curve (AUC) for the algorithms ranged from 0.556 to 0.994. The top-performing algorithm achieved a lesion-level, true-positive fraction comparable with that of the pathologist WOTC (72.4% [95% CI, 64.3%-80.4%]) at a mean of 0.0125 false-positives per normal whole-slide image. For the whole-slide image classification task, the best algorithm (AUC, 0.994 [95% CI, 0.983-0.999]) performed significantly better than the pathologists WTC in a diagnostic simulation (mean AUC, 0.810 [range, 0.738-0.884]; P < .001). The top 5 algorithms had a mean AUC that was comparable with the pathologist interpreting the slides in the absence of time constraints (mean AUC, 0.960 [range, 0.923-0.994] for the top 5 algorithms vs 0.966 [95% CI, 0.927-0.998] for the pathologist WOTC). Conclusions and Relevance: In the setting of a challenge competition, some deep learning algorithms achieved better diagnostic performance than a panel of 11 pathologists participating in a simulation exercise designed to mimic routine pathology workflow; algorithm performance was comparable with an expert pathologist interpreting whole-slide images without time constraints. Whether this approach has clinical utility will require evaluation in a clinical setting.