Obtaining spatially resolved tumor purity maps using deep multiple instance learning in a pan-cancer study

Is there a minimum percentage of sarcomatoid component required to affect outcomes of localised renal cell carcinoma?

OBJECTIVE

To evaluate and compare the outcomes of patients with localised renal cell carcinoma (RCC) with and without sarcomatoid features and the impact of this on cancer recurrence and survival.

MATERIAL AND METHODS

The Canadian Kidney Cancer information system database was used to identify patients diagnosed with localised RCC between January 2011 and December 2022. Patients with pT1-T3, n Nx-N0N1, M0 stage and documented sarcomatoid status were included. Patients with sarcomatoid RCC were categorised according to the sarcomatoid component percentage (%Sarc). Inverse probability of treatment weighting scores were used to balance the groups. Cox proportional hazards models were used to assess the impact of sarcomatoid status and %Sarc on recurrence-free and overall survival.

RESULTS

A total of 6660 patients (201 with and 6459 without sarcomatoid features) with non-metastatic RCC were included. %Sarc data were available in 155 patients, and the median value was 10%. The weighted analysis revealed that the presence of sarcomatoid features was associated with an increased risk of developing metastasis and increased risk of mortality compared to absence of sarcomatoid features. A %Sarc value >10 was associated with an increased risk of developing metastasis and of mortality compared to a %Sarc value ≤10.

CONCLUSIONS

Patients with a %Sarc >10 have an increased risk of recurrence and mortality. These patients may benefit from a more stringent follow-up and %Sarc could represent an important criterion in the risk assessment for adjuvant therapy.

A weakly supervised deep learning framework for automated PD-L1 expression analysis in lung cancer

The growing application of immune checkpoint inhibitors (ICIs) in cancer immunotherapy has underscored the critical need for reliable methods to identify patient populations likely to respond to ICI treatments, particularly in lung cancer treatment. Currently, the tumor proportion score (TPS), a crucial biomarker for patient selection, relies on manual interpretation by pathologists, which often shows substantial variability and inconsistency. To address these challenges, we innovatively developed multi-instance learning for TPS (MiLT), an innovative artificial intelligence (AI)-powered tool that predicts TPS from whole slide images. Our approach leverages multiple instance learning (MIL), which significantly reduces the need for labor-intensive cell-level annotations while maintaining high accuracy. In comprehensive validation studies, MiLT demonstrated remarkable consistency with pathologist assessments (intraclass correlation coefficient = 0.960, 95% confidence interval = 0.950-0.971) and robust performance across both internal and external cohorts. This tool not only standardizes TPS evaluation but also adapts to various clinical standards and provides time-efficient predictions, potentially transforming routine pathological practice. By offering a reliable, AI-assisted solution, MiLT could significantly improve patient selection for immunotherapy and reduce inter-observer variability among pathologists. These promising results warrant further exploration in prospective clinical trials and suggest new possibilities for integrating advanced AI in pathological diagnostics. MiLT represents a significant step toward more precise and efficient cancer immunotherapy decision-making.

Brain tumour histopathology through the lens of deep learning: A systematic review

PROBLEM

Machine learning (ML)/Deep learning (DL) techniques have been evolving to solve more complex diseases, but it has been used relatively little in Glioblastoma (GBM) histopathological studies, which could benefit greatly due to the disease's complex pathogenesis.

AIM

Conduct a systematic review to investigate how ML/DL techniques have influenced the progression of brain tumour histopathological research, particularly in GBM.

METHODS

54 eligible studies were collected from the PubMed and ScienceDirect databases, and their information about the types of brain tumour/s used, types of -omics data used with histopathological data, origins of the data, types of ML/DL and its training and evaluation methodologies, and the ML/DL task it was set to perform in the study were extracted to inform us of trends in GBM-related ML/DL-based research.

RESULTS

Only 8 GBM-related studies in the eligible utilised ML/DL methodologies to gain deeper insights into GBM pathogenesis by contextualising histological data with -omics data. However, we report that these studies have been published more recently. The most popular ML/DL models used in GBM-related research are the SVM classifier and ResNet-based CNN architecture. Still, a considerable number of studies failed to state training and evaluative methodologies clearly.

CONCLUSION

There is a growing trend towards using ML/DL approaches to uncover relationships between biological and histopathological data to bring new insights into GBM, thus pushing GBM research forward. Much work still needs to be done to properly report the ML/DL methodologies to showcase the models' robustness and generalizability and ensure the models are reproducible.

Tumour purity assessment with deep learning in colorectal cancer and impact on molecular analysis

Tumour content plays a pivotal role in directing the bioinformatic analysis of molecular profiles such as copy number variation (CNV). In clinical application, tumour purity estimation (TPE) is achieved either through visual pathological review [conventional pathology (CP)] or the deconvolution of molecular data. While CP provides a direct measurement, it demonstrates modest reproducibility and lacks standardisation. Conversely, deconvolution methods offer an indirect assessment with uncertain accuracy, underscoring the necessity for innovative approaches. SoftCTM is an open-source, multiorgan deep-learning (DL) model for the detection of tumour and non-tumour cells in H&E-stained slides, developed within the Overlapped Cell on Tissue Dataset for Histopathology (OCELOT) Challenge 2023. Here, using three large multicentre colorectal cancer (CRC) cohorts (N = 1,097 patients) with digital pathology and multi-omic data, we compare the utility and accuracy of TPE with SoftCTM versus CP and bioinformatic deconvolution methods (RNA expression, DNA methylation) for downstream molecular analysis, including CNV profiling. SoftCTM showed technical repeatability when applied twice on the same slide (r = 1.0) and excellent correlations in paired H&E slides (r > 0.9). TPEs profiled by SoftCTM correlated highly with RNA expression (r = 0.59) and DNA methylation (r = 0.40), while TPEs by CP showed a lower correlation with RNA expression (r = 0.41) and DNA methylation (r = 0.29). We show that CP and deconvolution methods respectively underestimate and overestimate tumour content compared to SoftCTM, resulting in 6-13% differing CNV calls. In summary, TPE with SoftCTM enables reproducibility, automation, and standardisation at single-cell resolution. SoftCTM estimates (M = 58.9%, SD ±16.3%) reconcile the overestimation by molecular data extrapolation (RNA expression: M = 79.2%, SD ±10.5, DNA methylation: M = 62.7%, SD ±11.8%) and underestimation by CP (M = 35.9%, SD ±13.1%), providing a more reliable middle ground. A fully integrated computational pathology solution could therefore be used to improve downstream molecular analyses for research and clinics. © 2024 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Multiple instance learning for digital pathology: A review of the state-of-the-art, limitations & future potential

Digital whole slides images contain an enormous amount of information providing a strong motivation for the development of automated image analysis tools. Particularly deep neural networks show high potential with respect to various tasks in the field of digital pathology. However, a limitation is given by the fact that typical deep learning algorithms require (manual) annotations in addition to the large amounts of image data, to enable effective training. Multiple instance learning exhibits a powerful tool for training deep neural networks in a scenario without fully annotated data. These methods are particularly effective in the domain of digital pathology, due to the fact that labels for whole slide images are often captured routinely, whereas labels for patches, regions, or pixels are not. This potential resulted in a considerable number of publications, with the vast majority published in the last four years. Besides the availability of digitized data and a high motivation from the medical perspective, the availability of powerful graphics processing units exhibits an accelerator in this field. In this paper, we provide an overview of widely and effectively used concepts of (deep) multiple instance learning approaches and recent advancements. We also critically discuss remaining challenges as well as future potential.

Distribution based MIL pooling filters: Experiments on a lymph node metastases dataset

Histopathology is a crucial diagnostic tool in cancer and involves the analysis of gigapixel slides. Multiple instance learning (MIL) promises success in digital histopathology thanks to its ability to handle gigapixel slides and work with weak labels. MIL is a machine learning paradigm that learns the mapping between bags of instances and bag labels. It represents a slide as a bag of patches and uses the slide's weak label as the bag's label. This paper introduces distribution-based pooling filters that obtain a bag-level representation by estimating marginal distributions of instance features. We formally prove that the distribution-based pooling filters are more expressive than the classical point estimate-based counterparts, like 'max' and 'mean' pooling, in terms of the amount of information captured while obtaining bag-level representations. Moreover, we empirically show that models with distribution-based pooling filters perform equal to or better than those with point estimate-based pooling filters on distinct real-world MIL tasks defined on the CAMELYON16 lymph node metastases dataset. Our model with a distribution pooling filter achieves an area under the receiver operating characteristics curve value of 0.9325 (95% confidence interval: 0.8798 - 0.9743) in the tumor vs. normal slide classification task.

Oner, Sung, and Lee: Researchers in digital pathology for the future of modern medicine

Oner, an early-career researcher, and Lee and Sung, group leaders, have developed a deep learning model for accurate prediction of the proportion of cancer cells within tumor tissue. This is a necessary step for precision oncology and target therapy in cancer. They talk about their view of data science and the evolution of pathology in the coming years.