A Soft Label Deep Learning to Assist Breast Cancer Target Therapy and Thyroid Cancer Diagnosis

Abstract

According to the World Health Organization Report 2022, cancer is the most common cause of death contributing to nearly one out of six deaths worldwide. Early cancer diagnosis and prognosis have become essential in reducing the mortality rate. On the other hand, cancer detection is a challenging task in cancer pathology. Trained pathologists can detect cancer, but their decisions are subjective to high intra- and inter-observer variability, which can lead to poor patient care owing to false-positive and false-negative results. In this study, we present a soft label fully convolutional network (SL-FCN) to assist in breast cancer target therapy and thyroid cancer diagnosis, using four datasets. To aid in breast cancer target therapy, the proposed method automatically segments human epidermal growth factor receptor 2 (HER2) amplification in fluorescence in situ hybridization (FISH) and dual in situ hybridization (DISH) images. To help in thyroid cancer diagnosis, the proposed method automatically segments papillary thyroid carcinoma (PTC) on Papanicolaou-stained fine needle aspiration and thin prep whole slide images (WSIs). In the evaluation of segmentation of HER2 amplification in FISH and DISH images, we compare the proposed method with thirteen deep learning approaches, including U-Net, U-Net with InceptionV5, Ensemble of U-Net with Inception-v4, Inception-Resnet-v2 encoder, and ResNet-34 encoder, SegNet, FCN, modified FCN, YOLOv5, CPN, SOLOv2, BCNet, and DeepLabv3+ with three different backbones, including MobileNet, ResNet, and Xception, on three clinical datasets, including two DISH datasets on two different magnification levels and a FISH dataset. The result on DISH breast dataset 1 shows that the proposed method achieves high accuracy of 87.77 ± 14.97%, recall of 91.20 ± 7.72%, and F1-score of 81.67 ± 17.76%, while, on DISH breast dataset 2, the proposed method achieves high accuracy of 94.64 ± 2.23%, recall of 83.78 ± 6.42%, and F1-score of 85.14 ± 6.61% and, on the FISH breast dataset, the proposed method achieves high accuracy of 93.54 ± 5.24%, recall of 83.52 ± 13.15%, and F1-score of 86.98 ± 9.85%, respectively. Furthermore, the proposed method outperforms most of the benchmark approaches by a significant margin (p <0.001). In evaluation of segmentation of PTC on Papanicolaou-stained WSIs, the proposed method is compared with three deep learning methods, including Modified FCN, U-Net, and SegNet. The experimental result demonstrates that the proposed method achieves high accuracy of 99.99 ± 0.01%, precision of 92.02 ± 16.6%, recall of 90.90 ± 14.25%, and F1-score of 89.82 ± 14.92% and significantly outperforms the baseline methods, including U-Net and FCN (p<0.001). With the high degree of accuracy, precision, and recall, the results show that the proposed method could be used in assisting breast cancer target therapy and thyroid cancer diagnosis with faster evaluation and minimizing human judgment errors.

Efficient Convolution Network to Assist Breast Cancer Diagnosis and Target Therapy

“…Raza et al [38] proposed Micro-Net which is a fully convolutional deep learning framework for segmentation of cells, nuclei and glands in microscopic images. In this study, we present an improved and extended DSL-FCN2s deep learning model that achieves almost similar results as the previous effort [8] but takes less time and memory usage for training and inference for practical clinical usage. Here, we develop a proposed method and compare it with thirteen baseline deep learning methods, including FCN [23], Modified FCN [4][5][6][7]9], SL-FCN [8], U-Net [24] + InceptionV4 [25], Ensemble of U-net with Inception-v4 [25], Inception-Resnet-v2 encoder [25], and ResNet-34 encoder [26], U-Net [24], SegNet [27], YOLOv5 [39], BCNet [28], CPN [29], SOLOv2 [30] and DeepLabv3+ [31] with three different backbones, including MobileNet [32], ResNet [26] and Xception [33].…”

“…When applied for pathological images, deep learning methods extract useful characteristics from pathological images, resulting in better diagnosis and patient outcomes. Although, our previous efforts using deep learning have yielded promising results in applications to segmentation of cervical cancer [4], breast cancer [5], ovarian cancer [6,7] and HER2 status evaluation in breast cancer [8], some challenges limit its utility in practice. Firstly and most importantly, large computational cost of deep learning is the primary barrier in deploying these models in routine clinical practice.…”

“…Secondly, when dealing with tissue or cell regions of interest, containing blurry or unclear boundaries, the performance of deep learning model deteriorates. To deal with the abovementioned issues, we propose an improved fully convolution network with low computing cost as an extended work of our previous studies [4][5][6][7][8][9], with three main improvements. Firstly, a dilated convolution is integrated into the proposed network to enlarge receptive fields for extracting multi-scale contextual information without losing resolution and greatly speeding up the model efficiently, hence reducing processing time and memory usage.…”

“…In Section 3, the materials and method are reported in details. In Section 4, we conduct a comparison study of the proposed method with thirteen recently published deep learning models, including FCN [23], Modified FCN [4][5][6][7]9], SL-FCN [8], U-Net [24] + In-ceptionV4 [25], Ensemble of U-net with Inception-v4 [25], Inception-Resnet-v2 encoder [25], and ResNet-34 encoder [26], U-Net [24], SegNet [27], BCNet [28], CPN [29], SOLOv2 [30] and DeepLabv3+ [31] with three different backbones, including MobileNet [32], ResNet [26] and Xception [33]. In Section 5, we further conducted run time analysis.…”

AI in Thyroid Cancer Diagnosis: Techniques, Trends, and Future Directions

“…• DL models' hyperparameters: Choosing the right DL algorithm is crucial in addressing various issues, particularly those related to thyroid cancer diagnosis. Due to the close similarities between benign and malignant tumors, as well as between tumors and other types of lymphocytes, it is challenging to differentiate between them accurately [268]. To achieve this, a significant increase in the number of layers for feature extraction may be required.…”