The diagnosis of primary liver cancers (PLCs) can be challenging, especially on biopsies and for combined hepatocellular-cholangiocarcinoma (cHCC-CCA). We automatically classified PLCs on routine-stained biopsies using a weakly supervised learning method. We selected 166 PLC biopsies divided into training, internal and external validation sets: 90, 29 and 47 samples, respectively. Two liver pathologists reviewed each whole-slide hematein eosin saffron (HES)-stained image (WSI). After annotating the tumour/non-tumour areas, tiles of 256x256 pixels were extracted from the WSIs and used to train a ResNet18 neural network. The tumour/non-tumour annotations served as labels during training, and the network's last convolutional layer was used to extract new tumour tile features. Without knowledge of the precise labels of the malignancies, we then applied an unsupervised clustering algorithm. Pathological review classified the training and validation sets into hepatocellular carcinoma (HCC, 33/90, 11/29 and 26/47), intrahepatic cholangiocarcinoma (iCCA, 28/90, 9/29 and 15/47), and cHCC-CCA (29/90, 9/29 and 6/47). In the two-cluster model, Clusters 0 and 1 contained mainly HCC and iCCA histological features. The diagnostic agreement between the pathological diagnosis and the two-cluster model predictions (major contingent) in the internal and external validation sets was 100% (11/11) and 96% (25/26) for HCC and 78% (7/9) and 87% (13/15) for iCCA, respectively. For cHCC-CCA, we observed a highly variable proportion of tiles from each cluster (cluster 0: 5-97%; cluster 1: 2-94%). Our method applied to PLC HES biopsy could identify specific morphological features of HCC and iCCA. Although no specific features of cHCC-CCA were recognized, assessing the proportion of HCC and iCCA tiles within a slide could facilitate the identification of cHCC-CCA. The diagnosis of primary liver cancers can be challenging, especially on biopsies and for combined hepatocellular-cholangiocarcinoma (cHCC-CCA). We automatically classified primary liver cancers on routine-stained biopsies using a weakly supervised learning method. Our model identified specific features of hepatocellular carcinoma and intrahepatic cholangiocarcinoma. Despite no specific features of cHCC-CCA being recognized, the identification of hepatocellular carcinoma and intrahepatic cholangiocarcinoma tiles within a slide could facilitate the diagnosis of primary liver cancers, and particularly cHCC-CCA.

Objectives: To share our experience with digital slide telepathology for intraoperative frozen section consultations (IOCs) and to describe its evolution over time by reporting performance metrics and addressing organizational and economic aspects.Methods: Since 2013, a technician has been alone at the surgical site. At the other site, the pathologist opens the digital slide from a local server via the intranet. Three periods were compared: a 6-month period of conventional IOC (period 1), a 24-month period of telepathology at 6 months after implementation (period 2), and a 12-month period of telepathology at 3.5 years after implementation (period 3).Results: In total, 87 conventional IOCs and 464 and 313 IOCs on digital slides were performed respectively during periods 1, 2, and 3; mean turnaround time was 27, 36, and 38 minutes, respectively, and there were a mean number of 1.1, 1.1, and 1.3 slides, respectively, per IOC. Diagnostic accuracy was achieved in 95.4%, 92.7%, and 93.9%, respectively, of IOCs (not significant). The additional cost is in the same range as the cost of urgent transport by courier.Conclusions: Developing IOC with digital slides is a challenge but is necessary to optimize medical time in the current context of pathologist shortage and budget restrictions.

In this article, we describe the case of a 34-year-old woman presenting a multifocal and metastatic epithelioid hemangioendothelioma (HEHE) of the liver. Under classical chemotherapy using cyclophosphamide, there was a fast tumor progression in liver and extra-hepatic metastatic sites (lungs and mediastinal lymph node). Taking into account the patient’s age and the natural history of the HEHE, our goal was to try to bring her to liver transplantation (LT) and lenvatinib was an acceptable candidate for this reason. Shortly after the initiation of lenvatinib before LT and surgery, we observed the enlargement of large devascularized necrotic areas in most of the liver HEHE masses, suggesting a good response. The patient was finally transplanted 6 months after initiation of lenvatinib treatment. Eight months after LT, progression occurred (ascites, peritoneal recurrence, and mediastinal lymph node). After restarting lenvatinib, ascites disappeared and the lymph node decreased in size, suggesting a good response, more than 1 year after her transplantation. This is the first case report to our knowledge that illustrates the benefit of lenvatinib as a neoadjuvant bridge until LT for a multifocal and metastatic HEHE. In addition, this drug has also shown a benefit in term of disease control after a late recurrence of the tumor. We suggest that lenvatinib should be proposed as a bridge to the LT for nonresectable HEHE. Moreover, this drug was also beneficial in the treatment of late recurrence after LT. The absence of pharmacologic interactions between classical immunosuppressive drugs and lenvatinib may allow its use as an early adjuvant approach when the risk of recurrence is high. The strength of our case consists in the long follow-up and the innovative message allowing changing palliative strategies into curative ones in case of advanced HEHE.