Deep learning is expected to aid pathologists by automating tasks such as tumour segmentation. We aimed to develop one universal tumour segmentation model for histopathological images and examine its performance in different cancer types. The model was developed using over 20 000 whole-slide images from over 4 000 patients with colorectal, endometrial, lung, or prostate carcinoma. Performance was validated in pre-planned analyses on external cohorts with over 3 000 patients across six cancer types. Exploratory analyses included over 1 500 additional patients from The Cancer Genome Atlas. Average Dice coefficient was over 80% in all validation cohorts with en bloc resection specimens and in The Cancer Genome Atlas cohorts. No loss of performance was observed when comparing the universal model with models specialised on single cancer types. In conclusion, extensive and rigorous evaluations demonstrate that generic tumour segmentation by a single model is possible across cancer types, patient populations, sample preparations, and slide scanners.

Logical assessment formula (LAF) [1] was proposed to achieve evaluations with inaccurate ground-truth labels (IAGTLs), which alleviates the usual evaluations with accurate ground-truth labels (AGTLs) [2-6], to assess predictive models for various artificial intelligence applications. LAF is suitable for evaluating the predicted targets of a predictive model in the situation, where the underlying true targets are difficult to be precisely defined while multiple inaccurate targets that contain various information consistent with our prior knowledge about the underlying true target are available. Theoretical analysis of LAF revealed the practicability of LAF for evaluations with IAGTLs, which includes: 1) LAF can be applied for evaluations with IAGTLs on a more difficult task, able to act like usual strategies for evaluations with AGTLs reasonably; and 2) LAF can be applied for evaluations with IAGTLs simply from the logical point of view on an easier task, unable to act like usual strategies for evaluations with AGTLs confidently. However, the revealed practicability of LAF for evaluations with IAGTLs has not yet been validated in real-world practice. In this paper, we aimed to address this issue. We applied LAF to tumour segmentation for breast cancer (TSfBC) in medical histopathology whole slide image analysis (MHWSIA). Extensive experiments were conducted and corresponding results and analyses support that the practicability of LAF is valid in the case of TSfBC in MHWSIA, which reflect the potentials of LAF applied to MHWSIA for evaluations with IAGTLs. The rest contents of this paper are structured as follows.

The complex heterogeneity of brain tumours is increasingly recognized to demand data of magnitudes and richness only fully-inclusive, large-scale collections drawn from routine clinical care could plausibly offer. This is a task contemporary machine learning could facilitate, especially in neuroimaging, but its ability to deal with incomplete data common in real world clinical practice remains unknown. Here we apply state-of-the-art methods to large scale, multi-site MRI data to quantify the comparative fidelity of automated tumour segmentation models replicating the various levels of sequence availability observed in the clinical reality. We compare deep learning (nnU-Net-derived) segmentation models with all possible combinations of T1, contrast-enhanced T1, T2, and FLAIR sequences, trained and validated with five-fold cross-validation on the 2021 BraTS-RSNA glioma population of 1251 patients, with further testing on a real-world 50 patient sample diverse in not only MRI scanner and field strength, but a random selection of pre- and post-operative imaging also. Models trained on incomplete imaging data segmented lesions well, often equivalently to those trained on complete data, exhibiting Dice coefficients of 0.907 (single sequence) to 0.945 (full datasets) for whole tumours, and 0.701 (single sequence) to 0.891 (full datasets) for component tissue types. Incomplete data segmentation models could accurately detect enhancing tumour in the absence of contrast imaging, quantifying its volume with an R2 between 0.95-0.97, and were invariant to lesion morphometry. Deep learning segmentation models characterize tumours well when missing data and can even detect enhancing tissue without the use of contrast. This suggests translation to clinical practice, where incomplete data is common, may be easier than hitherto believed, and may be of value in reducing dependence on contrast use.

Vestibular Schwannoma (VS) typically grows from the inner ear to the brain. It can be separated into two regions, intrameatal and extrameatal respectively corresponding to being inside or outside the inner ear canal. The growth of the extrameatal regions is a key factor that determines the disease management followed by the clinicians. In this work, a VS segmentation approach with subdivision into intra-/extra-meatal parts is presented. We annotated a dataset consisting of 227 T2 MRI instances, acquired longitudinally on 137 patients, excluding post-operative instances. We propose a staged approach, with the first stage performing the whole tumour segmentation and the second stage performing the intra-/extra-meatal segmentation using the T2 MRI along with the mask obtained from the first stage. To improve on the accuracy of the predicted meatal boundary, we introduce a task-specific loss which we call Boundary Distance Loss. The performance is evaluated in contrast to the direct intrameatal extrameatal segmentation task performance, i.e. the Baseline. Our proposed method, with the two-stage approach and the Boundary Distance Loss, achieved a Dice score of 0.8279+-0.2050 and 0.7744+-0.1352 for extrameatal and intrameatal regions respectively, significantly improving over the Baseline, which gave Dice score of 0.7939+-0.2325 and 0.7475+-0.1346 for the extrameatal and intrameatal regions respectively.

What is Artificial Intelligence (AI) and what are the problems it can solve in healthcare? "[The automation of] activities that we associate with human thinking, activities such as decision-making, problem solving, learning …" Machines have far superior computational abilities than humans. They can sort through enormous amounts of data and use it to make better decisions. What are the main components of AI? AI has numerous application areas where it is presenting ground-breaking human level results. All these areas are evolving on daily basis, new research innovations are occurring.

Logical assessment formula (LAF) [1] was proposed to alleviate the situation where accurate ground-truth labels are not available for evaluations of an approach for a specific learning task. With a H. pylori segmentation task of medical histopathology whole slide images [1,2], evaluations based on LAF has been qualitatively shown to be able to reflect the logical rationalities of the predictions of various approaches. Comprehensive descriptions of LAF can be found in Section 2. However, the principles of LAF for evaluations without accurate ground-truth labels (AGTL) are not well revealed. In this paper, we provide comprehensive theoretical analyses to reveal the principles of LAF for evaluations without AGTL. Details of the revealed principles of LAF are presented in Section 4. From the revealed principles of LAF, we summarize, for the situation where accurate ground-truth labels are not available while multiple inaccurate targets containing various information consistent with our prior knowledge about the true target are available, the major practicability of LAF is that it can be reasonably applied for evaluations without AGTL on a more difficult task, just acting like usual strategies for evaluations with AGTL; and the minor practicability of LAF is that it can be applied for evaluations without AGTL from the logical perspective on an easier task, unable to be acting like usual strategies for evaluations with AGTL. Details of the practicability of LAF summarized from the revealed principles can be found in Section 5. To verify the practicability of LAF summarized from the revealed principles, we apply LAF on two tumour segmentation tasks in medical histopathology whole slide images (MHWSI) for breast cancer for evaluations without AGTL. Experimental results analyses of LAF applied on tumour segmentation for breast cancer support the practicability of LAF summarized from the revealed principles. Comprehensive contents can be found in Section 6 and 7.

Accurate brain tumour segmentation is a crucial step towards improving disease diagnosis and proper treatment planning. In this paper, we propose a deep-learning based method to segment a brain tumour into its subregions: whole tumour, tumour core and enhancing tumour. The proposed architecture is a 3D convolutional neural network based on a variant of the U-Net architecture of Ronneberger et al. [17] with three main modifications: (i) a heavy encoder, light decoder structure using residual blocks (ii) employment of dense blocks instead of skip connections, and (iii) utilization of self-ensembling in the decoder part of the network. The network was trained and tested using two different approaches: a multitask framework to segment all tumour subregions at the same time, and a three-stage cascaded framework to segment one subregion at a time. An ensemble of the results from both frameworks was also computed. To address the class imbalance issue, appropriate patch extraction was employed in a pre-processing step. Connected component analysis was utilized in the post-processing step to reduce the false positive predictions. Experimental results on the BraTS20 validation dataset demonstrates that the proposed model achieved average Dice Scores of 0.90, 0.82, and 0.78 for whole tumour, tumour core and enhancing tumour respectively. Keywords: Brain tumour segmentation, · Multimodal MRI, · Cascaded network, · Densely connected CNN.