Deep learning for digital pathology is hindered by the extremely high spatial resolution of whole-slide images (WSIs). Most studies have employed patch-based methods, which often require detailed annotation of image patches. This typically involves laborious free-hand contouring on WSIs. To alleviate the burden of such contouring and obtain benefits from scaling up training with numerous WSIs, we develop a method for training neural networks on entire WSIs using only slide-level diagnoses. Our method leverages the unified memory mechanism to overcome the memory constraint of compute accelerators. Experiments conducted on a data set of 9662 lung cancer WSIs reveal that the proposed method achieves areas under the receiver operating characteristic curve of 0.9594 and 0.9414 for adenocarcinoma and squamous cell carcinoma classification on the testing set, respectively. Furthermore, the method demonstrates higher classification performance than multiple-instance learning as well as strong localization results for small lesions through class activation mapping.

Human spinal balance assessment relies considerably on sagittal radiographic parameter measurement. Deep learning could be applied for automatic landmark detection and alignment analysis, with mild to moderate standard errors and favourable correlations with manual measurement. In this study, based on 2210 annotated images of various spinal disease aetiologies, we developed deep learning models capable of automatically locating 45 anatomic landmarks and subsequently generating 18 radiographic parameters on a whole-spine lateral radiograph. In the assessment of model performance, the localisation accuracy and learning speed were the highest for landmarks in the cervical area, followed by those in the lumbosacral, thoracic, and femoral areas. All the predicted radiographic parameters were significantly correlated with ground truth values (all p < 0.001). The human and artificial intelligence comparison revealed that the deep learning model was capable of matching the reliability of doctors for 15/18 of the parameters. The proposed automatic alignment analysis system was able to localise spinal anatomic landmarks with high accuracy and to generate various radiographic parameters with favourable correlations with manual measurements.

The pathological identification of lymph node (LN) metastasis is demanding and tedious. Although convolutional neural networks (CNNs) possess considerable potential in improving the process, the ultrahigh-resolution of whole slide images hinders the development of a clinically applicable solution. We design an artificial-intelligence-assisted LN assessment workflow to facilitate the routine counting of metastatic LNs. Unlike previous patch-based approaches, our proposed method trains CNNs by using 5-gigapixel images, obviating the need for lesion-level annotations. Trained on 5907 LN images, our algorithm identifies metastatic LNs in gastric cancer with a slide-level area under the receiver operating characteristic curve (AUC) of 0.9936. Clinical experiments reveal that the workflow significantly improves the sensitivity of micrometastasis identification (81.94% to 95.83%, P < .001) and isolated tumor cells (67.95% to 96.15%, P < .001) in a significantly shorter review time (−31.5%, P < .001). Cross-site evaluation indicates that the algorithm is highly robust (AUC = 0.9829).

Detection of nodal micrometastasis (tumor size: 0.2-2.0 mm) is challenging for pathologists due to the small size of metastatic foci. Since lymph nodes with micrometastasis are counted as positive nodes, detecting micrometastasis is crucial for accurate pathologic staging of colorectal cancer. Previously, deep learning algorithms developed with manually annotated images performed well in identifying micrometastasis of breast cancer in sentinel lymph nodes. However, the process of manual annotation is labor intensive and time consuming. Multiple instance learning was later used to identify metastatic breast cancer without manual annotation, but its performance appears worse in detecting micrometastasis. Here, we developed a deep learning model using whole-slide images of regional lymph nodes of colorectal cancer with only a slide-level label (either a positive or negative slide). The training, validation, and testing sets included 1963, 219, and 1000 slides, respectively. A supercomputer TAIWANIA 2 was used to train a deep learning model to identify metastasis. At slide level, our algorithm performed well in identifying both macrometastasis (tumor size > 2.0 mm) and micrometastasis with an area under the receiver operating characteristics curve (AUC) of 0.9993 and 0.9956, respectively. Since most of our slides had more than one lymph node, we then tested the performance of our algorithm on 538 single-lymph node images randomly cropped from the testing set. At single-lymph node level, our algorithm maintained good performance in identifying macrometastasis and micrometastasis with an AUC of 0.9944 and 0.9476, respectively. Visualization using class activation mapping confirmed that our model identified nodal metastasis based on areas of tumor cells. Our results demonstrate for the first time that micrometastasis could be detected by deep learning on whole-slide images without manual annotation.

Background: Rosacea is a chronic inflammatory disease with variable clinical presentations, including transient flushing, fixed erythema, papules, pustules, and phymatous changes on the central face. Owing to the diversity in the clinical manifestations of rosacea, the lack of objective biochemical examinations, and nonspecificity in histopathological findings, accurate identification of rosacea is a big challenge. Artificial intelligence has emerged as a potential tool in the identification and evaluation of some skin diseases such as melanoma, basal cell carcinoma, and psoriasis. Objective: The objective of our study was to utilize a convolutional neural network (CNN) to differentiate the clinical photos of patients with rosacea (taken from 3 different angles) from those of patients with other skin diseases such as acne, seborrheic dermatitis, and eczema that could be easily confused with rosacea. Methods: In this study, 24,736 photos comprising of 18,647 photos of patients with rosacea and 6089 photos of patients with other skin diseases such as acne, facial seborrheic dermatitis, and eczema were included and analyzed by our CNN model based on ResNet-50. Results: The CNN in our study achieved an overall accuracy and precision of 0.914 and 0.898, with an area under the receiver operating characteristic curve of 0.972 for the detection of rosacea. The accuracy of classifying 3 subtypes of rosacea, that is, erythematotelangiectatic rosacea, papulopustular rosacea, and phymatous rosacea was 83.9%, 74.3%, and 80.0%, respectively. Moreover, the accuracy and precision of our CNN to distinguish rosacea from acne reached 0.931 and 0.893, respectively. For the differentiation between rosacea, seborrheic dermatitis, and eczema, the overall accuracy of our CNN was 0.757 and the precision was 0.667. Finally, by comparing the CNN diagnosis with the diagnoses by dermatologists of different expertise levels, we found that our CNN system is capable of identifying rosacea with a performance superior to that of resident doctors or attending physicians and comparable to that of experienced dermatologists. Conclusions: The findings of our study showed that by assessing clinical images, the CNN system in our study could identify rosacea with accuracy and precision comparable to that of an experienced dermatologist.

Pathologic diagnosis of nasopharyngeal carcinoma (NPC) can be challenging since most cases are nonkeratinizing carcinoma with little differentiation and many admixed lymphocytes. Our aim was to evaluate the possibility to identify NPC in nasopharyngeal biopsies using deep learning. A total of 726 nasopharyngeal biopsies were included. Among them, 100 cases were randomly selected as the testing set, 20 cases as the validation set, and all other 606 cases as the training set. All three datasets had equal numbers of NPC cases and benign cases. Manual annotation was performed. Cropped square image patches of 256 × 256 pixels were used for patch-level training, validation, and testing. The final patch-level algorithm effectively identified NPC patches, with an area under the receiver operator characteristic curve (AUC) of 0.9900. Using gradient-weighted class activation mapping, we demonstrated that the identification of NPC patches was based on morphologic features of tumor cells. At the second stage, whole-slide images were sequentially cropped into patches, inferred with the patch-level algorithm, and reconstructed into images with a smaller size for training, validation, and testing. Finally, the AUC was 0.9848 for slide-level identification of NPC. Our result shows for the first time that deep learning algorithms can identify NPC.