Digital Pathology: Basics, Clinical Applications and Future Trends

Diagnosing primary liver cancers, particularly hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC), is a challenging and labor-intensive process, even for experts, and secondary liver cancers further complicate the diagnosis. Artificial intelligence (AI) offers promising solutions to these diagnostic challenges by facilitating the histopathological classification of tumors using digital whole slide images (WSIs). This study aimed to develop a deep learning model for distinguishing HCC, CC, and metastatic colorectal cancer (mCRC) using histopathological images and to discuss its clinical implications. The WSIs from HCC, CC, and mCRC were used to train the classifiers. For normal/tumor classification, the areas under the curve (AUCs) were 0.989, 0.988, and 0.991 for HCC, CC, and mCRC, respectively. Using proper tumor tissues, the HCC/other cancer type classifier was trained to effectively distinguish HCC from CC and mCRC, with a concatenated AUC of 0.998. Subsequently, the CC/mCRC classifier differentiated CC from mCRC with a concatenated AUC of 0.995. However, testing on an external dataset revealed that the HCC/other cancer type classifier underperformed with an AUC of 0.745. After combining the original training datasets with external datasets and retraining, the classification drastically improved, all achieving AUCs of 1.000. Although these results are promising and offer crucial insights into liver cancer, further research is required for model refinement and validation. Full article

Lung cancer is the leading cause of cancer death in the U.S. Therefore, it is imperative to identify novel biomarkers for the early detection and progression of lung cancer. PRMT6 is associated with poor lung cancer prognosis. However, analyzing PRMT6 expression manually in large samples is time-consuming posing a significant limitation for processing this biomarker. To overcome this issue, we trained and validated an automated method for scoring PRMT6 in lung cancer tissues, which can then be used as the standard method in future larger cohorts to explore population-level associations between PRMT6 expression and sociodemographic/clinicopathologic characteristics. We evaluated the ability of a trained artificial intelligence (AI) algorithm to reproduce the PRMT6 immunoreactive scores obtained by pathologists. Our findings showed that tissue segmentation to cancer vs. non-cancer tissues was the most critical parameter, which required training and adjustment of the algorithm to prevent scoring non-cancer tissues or ignoring relevant cancer cells. The trained algorithm showed a high concordance with pathologists with a correlation coefficient of 0.88. The inter-rater agreement was significant, with an intraclass correlation of 0.95 and a scale reliability coefficient of 0.96. In conclusion, we successfully optimized a machine learning algorithm for scoring PRMT6 expression in lung cancer that matches the degree of accuracy of scoring by pathologists. Full article

The tumor–stroma ratio (TSR) has been repeatedly shown to be a prognostic factor for survival prediction of different cancer types. However, an objective and reliable determination of the tumor–stroma ratio remains challenging. We present an easily adaptable deep learning model for accurately segmenting tumor regions in hematoxylin and eosin (H&E)-stained whole slide images (WSIs) of colon cancer patients into five distinct classes (tumor, stroma, necrosis, mucus, and background). The tumor–stroma ratio can be determined in the presence of necrotic or mucinous areas. We employ a few-shot model, eventually aiming for the easy adaptability of our approach to related segmentation tasks or other primaries, and compare the results to a well-established state-of-the art approach (U-Net). Both models achieve similar results with an overall accuracy of 86.5% and 86.7%, respectively, indicating that the adaptability does not lead to a significant decrease in accuracy. Moreover, we comprehensively compare with TSR estimates of human observers and examine in detail discrepancies and inter-rater reliability. Adding a second survey for segmentation quality on top of a first survey for TSR estimation, we found that TSR estimations of human observers are not as reliable a ground truth as previously thought. Full article

Neuroblastoma is the most common extracranial, malignant, solid tumor found in children. In more than one-third of cases, the tumor is in an advanced stage, with limited resectability. The treatment options include resection, with or without (neo-/) adjuvant therapy, and conservative therapy, the latter even with curative intent. Contrast-enhanced MRI is used for staging and therapy monitoring. Diffusion-weighted imaging (DWI) is often included. DWI allows for a calculation of the apparent diffusion coefficient (ADC) for quantitative assessment. Histological tumor characteristics can be derived from ADC maps. Monitoring the response to treatment is possible using ADC maps, with an increase in ADC values in cases of a response to therapy. Changes in the ADC value precede volume reduction. The usual criteria for determining the response to therapy can therefore be supplemented by ADC values. While these changes have been observed in neuroblastoma, early changes in the ADC value in response to therapy are less well described. In this study, we evaluated whether there is an early change in the ADC values in neuroblastoma under therapy; if this change depends on the form of therapy; and whether this change may serve as a prognostic marker. We retrospectively evaluated neuroblastoma cases treated in our institution between June 2007 and August 2014. The examinations were grouped as ‘prestaging’; ‘intermediate staging’; ‘final staging’; and ‘follow-up’. A classification of “progress”, “stable disease”, or “regress” was made. For the determination of ADC values, regions of interest were drawn along the borders of all tumor manifestations. To calculate ADC changes (∆ADC), the respective MRI of the prestaging was used as a reference point or, in the case of therapies that took place directly after previous therapies, the associated previous staging. In the follow-up examinations, the previous examination was used as a reference point. The ∆ADC were grouped into ∆ADCregress for regressive disease, ∆ADCstable for stable disease, and ∆ADC for progressive disease. In addition, examinations at 60 to 120 days from the baseline were grouped as er∆ADCregress, er∆ADCstable, and er∆ADCprogress. Any differences were tested for significance using the Mann–Whitney test (level of significance: p < 0.05). In total, 34 patients with 40 evaluable tumor manifestations and 121 diffusion-weighted MRI examinations were finally included. Twenty-seven patients had INSS stage IV neuroblastoma, and seven had INSS stage III neuroblastoma. A positive N-Myc expression was found in 11 tumor diseases, and 17 patients tested negative for N-Myc (with six cases having no information). 26 patients were assigned to the high-risk group according to INRG and eight patients to the intermediate-risk group. There was a significant difference in mean ADC values from the high-risk group compared to those from the intermediate-risk group, according to INRG. The differences between the mean ∆ADC values (absolute and percentage) according to the course of the disease were significant: between ∆ADCregress and ∆ADCstable, between ∆ADCprogress and ∆ADCstable, as well as between ∆ADCregress and ∆ADCprogress. The differences between the mean er∆ADC values (absolute and percentage) according to the course of the disease were significant: between er∆ADCregress and er∆ADCstable, as well as between er∆ADCregress and er∆ADCprogress. Forms of therapy, N-Myc status, and risk groups showed no further significant differences in mean ADC values and ∆ADC/er∆ADC. A clear connection between the ADC changes and the response to therapy could be demonstrated. This held true even within the first 120 days after the start of therapy: an increase in the ADC value corresponds to a probable response to therapy, while a decrease predicts progression. Minimal or no changes were seen in cases of stable disease. Full article

Bladder cancer (BC) diagnosis and prediction of prognosis are hindered by subjective pathological evaluation, which may cause misdiagnosis and under-/over-treatment. Computational pathology (CPATH) can identify clinical outcome predictors, offering an objective approach to improve prognosis. However, a systematic review of CPATH in BC literature is lacking. Therefore, we present a comprehensive overview of studies that used CPATH in BC, analyzing 33 out of 2285 identified studies. Most studies analyzed regions of interest to distinguish normal versus tumor tissue and identify tumor grade/stage and tissue types (e.g., urothelium, stroma, and muscle). The cell’s nuclear area, shape irregularity, and roundness were the most promising markers to predict recurrence and survival based on selected regions of interest, with >80% accuracy. CPATH identified molecular subtypes by detecting features, e.g., papillary structures, hyperchromatic, and pleomorphic nuclei. Combining clinicopathological and image-derived features improved recurrence and survival prediction. However, due to the lack of outcome interpretability and independent test datasets, robustness and clinical applicability could not be ensured. The current literature demonstrates that CPATH holds the potential to improve BC diagnosis and prediction of prognosis. However, more robust, interpretable, accurate models and larger datasets—representative of clinical scenarios—are needed to address artificial intelligence’s reliability, robustness, and black box challenge. Full article