Search

Convolutional neural networks have demonstrated excellent performance in oral cancer detection and classification. However, the end-to-end learning strategy makes CNNs hard to interpret, and it can be challenging to fully understand the decision-making procedure. Additionally, reliability is also a significant challenge for CNN based approaches. In this study, we proposed a neural network called the attention branch network (ABN), which combines the visual explanation and attention mechanisms to improve the recognition performance and interpret the decision-making simultaneously. We also embedded expert knowledge into the network by having human experts manually edit the attention maps for the attention mechanism. Our experiments have shown that ABN performs better than the original baseline network. By introducing the Squeeze-and-Excitation (SE) blocks to the network, the cross-validation accuracy increased further. Furthermore, we observed that some previously misclassified cases were correctly recognized after updating by manually editing the attention maps. The cross-validation accuracy increased from 0.846 to 0.875 with the ABN (Resnet18 as baseline), 0.877 with SE-ABN, and 0.903 after embedding expert knowledge. The proposed method provides an accurate, interpretable, and reliable oral cancer computer-aided diagnosis system through visual explanation, attention mechanisms, and expert knowledge embedding.

SignificanceOral cancer is one of the most prevalent cancers, especially in middle- and low-income countries such as India. Automatic segmentation of oral cancer images can improve the diagnostic workflow, which is a significant task in oral cancer image analysis. Despite the remarkable success of deep-learning networks in medical segmentation, they rarely provide uncertainty quantification for their output.AimWe aim to estimate uncertainty in a deep-learning approach to semantic segmentation of oral cancer images and to improve the accuracy and reliability of predictions.ApproachThis work introduced a UNet-based Bayesian deep-learning (BDL) model to segment potentially malignant and malignant lesion areas in the oral cavity. The model can quantify uncertainty in predictions. We also developed an efficient model that increased the inference speed, which is almost six times smaller and two times faster (inference speed) than the original UNet. The dataset in this study was collected using our customized screening platform and was annotated by oral oncology specialists.ResultsThe proposed approach achieved good segmentation performance as well as good uncertainty estimation performance. In the experiments, we observed an improvement in pixel accuracy and mean intersection over union by removing uncertain pixels. This result reflects that the model provided less accurate predictions in uncertain areas that may need more attention and further inspection. The experiments also showed that with some performance compromises, the efficient model reduced computation time and model size, which expands the potential for implementation on portable devices used in resource-limited settings.ConclusionsOur study demonstrates the UNet-based BDL model not only can perform potentially malignant and malignant oral lesion segmentation, but also can provide informative pixel-level uncertainty estimation. With this extra uncertainty information, the accuracy and reliability of the model’s prediction can be improved.

Early detection of oral cancer in low-resource settings necessitates a Point-of-Care screening tool that empowers Frontline-Health-Workers (FHW). This study was conducted to validate the accuracy of Convolutional-Neural-Network (CNN) enabled m(mobile)-Health device deployed with FHWs for delineation of suspicious oral lesions (malignant/potentially-malignant disorders). The effectiveness of the device was tested in tertiary-care hospitals and low-resource settings in India. The subjects were screened independently, either by FHWs alone or along with specialists. All the subjects were also remotely evaluated by oral cancer specialist/s. The program screened 5025 subjects (Images: 32,128) with 95% (n = 4728) having telediagnosis. Among the 16% (n = 752) assessed by onsite specialists, 20% (n = 102) underwent biopsy. Simple and complex CNN were integrated into the mobile phone and cloud respectively. The onsite specialist diagnosis showed a high sensitivity (94%), when compared to histology, while telediagnosis showed high accuracy in comparison with onsite specialists (sensitivity: 95%; specificity: 84%). FHWs, however, when compared with telediagnosis, identified suspicious lesions with less sensitivity (60%). Phone integrated, CNN (MobileNet) accurately delineated lesions (n = 1416; sensitivity: 82%) and Cloud-based CNN (VGG19) had higher accuracy (sensitivity: 87%) with tele-diagnosis as reference standard. The results of the study suggest that an automated mHealth-enabled, dual-image system is a useful triaging tool and empowers FHWs for oral cancer screening in low-resource settings.

SignificanceConvolutional neural networks (CNNs) show the potential for automated classification of different cancer lesions. However, their lack of interpretability and explainability makes CNNs less than understandable. Furthermore, CNNs may incorrectly concentrate on other areas surrounding the salient object, rather than the network's attention focusing directly on the object to be recognized, as the network has no incentive to focus solely on the correct subjects to be detected. This inhibits the reliability of CNNs, especially for biomedical applications.AimDevelop a deep learning training approach that could provide understandability to its predictions and directly guide the network to concentrate its attention and accurately delineate cancerous regions of the image.ApproachWe utilized Selvaraju et al.'s gradient-weighted class activation mapping to inject interpretability and explainability into CNNs. We adopted a two-stage training process with data augmentation techniques and Li et al.'s guided attention inference network (GAIN) to train images captured using our customized mobile oral screening devices. The GAIN architecture consists of three streams of network training: classification stream, attention mining stream, and bounding box stream. By adopting the GAIN training architecture, we jointly optimized the classification and segmentation accuracy of our CNN by treating these attention maps as reliable priors to develop attention maps with more complete and accurate segmentation.ResultsThe network's attention map will help us to actively understand what the network is focusing on and looking at during its decision-making process. The results also show that the proposed method could guide the trained neural network to highlight and focus its attention on the correct lesion areas in the images when making a decision, rather than focusing its attention on relevant yet incorrect regions.ConclusionsWe demonstrate the effectiveness of our approach for more interpretable and reliable oral potentially malignant lesion and malignant lesion classification.

In medical imaging, deep learning-based solutions have achieved state-of-the-art performance. However, reliability restricts the integration of deep learning into practical medical workflows since conventional deep learning frameworks cannot quantitatively assess model uncertainty. In this work, we propose to address this shortcoming by utilizing a Bayesian deep network capable of estimating uncertainty to assess oral cancer image classification reliability. We evaluate the model using a large intraoral cheek mucosa image dataset captured using our customized device from high-risk population to show that meaningful uncertainty information can be produced. In addition, our experiments show improved accuracy by uncertainty-informed referral. The accuracy of retained data reaches roughly 90% when referring either 10% of all cases or referring cases whose uncertainty value is greater than 0.3. The performance can be further improved by referring more patients. The experiments show the model is capable of identifying difficult cases needing further inspection.

SignificanceEarly detection of oral cancer is vital for high-risk patients, and machine learning-based automatic classification is ideal for disease screening. However, current datasets collected from high-risk populations are unbalanced and often have detrimental effects on the performance of classification.AimTo reduce the class bias caused by data imbalance.ApproachWe collected 3851 polarized white light cheek mucosa images using our customized oral cancer screening device. We use weight balancing, data augmentation, undersampling, focal loss, and ensemble methods to improve the neural network performance of oral cancer image classification with the imbalanced multi-class datasets captured from high-risk populations during oral cancer screening in low-resource settings.ResultsBy applying both data-level and algorithm-level approaches to the deep learning training process, the performance of the minority classes, which were difficult to distinguish at the beginning, has been improved. The accuracy of "premalignancy" class is also increased, which is ideal for screening applications.ConclusionsExperimental results show that the class bias induced by imbalanced oral cancer image datasets could be reduced using both data- and algorithm-level methods. Our study may provide an important basis for helping understand the influence of unbalanced datasets on oral cancer deep learning classifiers and how to mitigate.

Non-invasive strategies that can identify oral malignant and dysplastic oral potentially-malignant lesions (OPML) are necessary in cancer screening and long-term surveillance. Optical coherence tomography (OCT) can be a rapid, real time and non-invasive imaging method for frequent patient surveillance. Here, we report the validation of a portable, robust OCT device in 232 patients (lesions: 347) in different clinical settings. The device deployed with algorithm-based automated diagnosis, showed efficacy in delineation of oral benign and normal (n = 151), OPML (n = 121), and malignant lesions (n = 75) in community and tertiary care settings. This study showed that OCT images analyzed by automated image processing algorithm could distinguish the dysplastic-OPML and malignant lesions with a sensitivity of 95% and 93%, respectively. Furthermore, we explored the ability of multiple (n = 14) artificial neural network (ANN) based feature extraction techniques for delineation high grade-OPML (moderate/severe dysplasia). The support vector machine (SVM) model built over ANN, delineated high-grade dysplasia with sensitivity of 83%, which in turn, can be employed to triage patients for tertiary care. The study provides evidence towards the utility of the robust and low-cost OCT instrument as a point-of-care device in resource-constrained settings and the potential clinical application of device in screening and surveillance of oral cancer.

SignificanceOral cancer is among the most common cancers globally, especially in low- and middle-income countries. Early detection is the most effective way to reduce the mortality rate. Deep learning-based cancer image classification models usually need to be hosted on a computing server. However, internet connection is unreliable for screening in low-resource settings.AimTo develop a mobile-based dual-mode image classification method and customized Android application for point-of-care oral cancer detection.ApproachThe dataset used in our study was captured among 5025 patients with our customized dual-modality mobile oral screening devices. We trained an efficient network MobileNet with focal loss and converted the model into TensorFlow Lite format. The finalized lite format model is ∼16.3  MB and ideal for smartphone platform operation. We have developed an Android smartphone application in an easy-to-use format that implements the mobile-based dual-modality image classification approach to distinguish oral potentially malignant and malignant images from normal/benign images.ResultsWe investigated the accuracy and running speed on a cost-effective smartphone computing platform. It takes ∼300  ms to process one image pair with the Moto G5 Android smartphone. We tested the proposed method on a standalone dataset and achieved 81% accuracy for distinguishing normal/benign lesions from clinically suspicious lesions, using a gold standard of clinical impression based on the review of images by oral specialists.ConclusionsOur study demonstrates the effectiveness of a mobile-based approach for oral cancer screening in low-resource settings.

ObjectivesSurgical margin status is a significant determinant of treatment outcome in oral cancer. Negative surgical margins can decrease the loco-regional recurrence by five-fold. The current standard of care of intraoperative clinical examination supplemented by histological frozen section, can result in a risk of positive margins from 5 to 17 percent. In this study, we attempted to assess the utility of intraoperative optical coherence tomography (OCT) imaging with automated diagnostic algorithm to improve on the current method of clinical evaluation of surgical margin in oral cancer.Materials and methodsWe have used a modified handheld OCT device with automated algorithm based diagnostic platform for imaging. Intraoperatively, images of 125 sites were captured from multiple zones around the tumor of oral cancer patients (n = 14) and compared with the clinical and pathologic diagnosis.ResultsOCT showed sensitivity and specificity of 100%, equivalent to histological diagnosis (kappa, ĸ = 0.922), in detection of malignancy within tumor and tumor margin areas. In comparison, for dysplastic lesions, OCT-based detection showed a sensitivity of 92.5% and specificity of 68.8% and a moderate concordance with histopathology diagnosis (ĸ = 0.59). Additionally, the OCT scores could significantly differentiate squamous cell carcinoma (SCC) from dysplastic lesions (mild/moderate/severe; p ≤ 0.005) as well as the latter from the non-dysplastic lesions (p ≤ 0.05).ConclusionThe current challenges associated with clinical examination-based margin assessment could be improved with intra-operative OCT imaging. OCT is capable of identifying microscopic tumor at the surgical margins and demonstrated the feasibility of mapping of field cancerization around the tumor.

Objectives: The incidence of young‐onset oral squamous cell carcinoma (OSCC) is growing, even among non‐smokers/drinkers. The effects of adverse histopathological features on long‐term oncologic outcomes between the young and old are controversial and confounded by significant heterogeneity. Few studies have evaluated the socio‐economic impact of premature mortality from OSCC. Our study seeks to quantify these differences and their economic impact on society. Materials and Methods: Four hundred and seventy‐eight young (<45 years) and 1660 old patients (≥45 years) with OSCC were studied. Logistic regression determined predictors of recurrence and death. Survival analysis was calculated via the Kaplan–Meier method. A separate health economic analysis was conducted for India and Singapore. Years of Potential Productive Life Lost (YPPLL) were estimated with the Human Capital Approach, and premature mortality cost was derived using population‐level data. Results: Adverse histopathological features were seen more frequently in young OSCC: PNI (42.9% vs. 35%, p = 0.002), LVI (22.4% vs. 17.3%, p = 0.013) and ENE (36% vs. 24.5%, p < 0.001). Although 5‐year OS/DSS were similar, the young cohort had received more intensive adjuvant therapy (CCRT 26.9% vs. 16.6%, p < 0.001). Among Singaporean males, the premature mortality cost per death was US $396,528, and per YPPLL was US $45,486. This was US $397,402 and US $38,458 for females. Among Indian males, the premature mortality cost per death was US $30,641, and per YPPLL was US $595. This was US $ 21,038 and US $305 for females. Conclusion: Young‐onset OSCC is an aggressive disease, mitigated by the ability to receive intensive adjuvant treatment. From our loss of productivity analysis, the socio‐economic costs from premature mortality are substantial. Early cancer screening and educational outreach campaigns should be tailored to this cohort. Alongside, more funding should be diverted to genetic research, developing novel biomarkers and improving the efficacy of adjuvant treatment in OSCC. [ABSTRACT FROM AUTHOR]

Background: Oral squamous cell carcinoma (OSCC) Is a common head and neck cancer with high morbidity and mortality. Currently, treatment decisions are guided by TNM staging, which omits important negative prognosticators such as lymphovascular invasion, perineural invasion (PNI), and histologic differentiation. We proposed nomogram models based on adverse pathological features to identify candidates suitable for treatment escalation within each risk group according to the National Comprehensive Cancer Network (NCCN) guidelines. Methods: Anonymized clinicopathologic data of OSCC patients from 5 tertiary healthcare institutions in Asia were divided into 3 risk groups according to the NCCN guidelines. Within each risk group, nomograms were built to predict overall survival based on histologic differentiation, histologic margin involvement, depth of invasion (DOI), extranodal extension, PNI, lymphovascular, and bone invasion. Nomograms were internally validated with precision-recall analysis and the Kaplan-Meier survival analysis. Results: Low-risk patients with positive pathological nodal involvement and/or positive PNI should be considered for adjuvant radiotherapy. Intermediate-risk patients with gross bone invasion may benefit from concurrent chemotherapy. High-risk patients with positive margins, high DOI, and a high composite score of histologic differentiation, PNI, and the American Joint Committee on Cancer (AJCC) 8th edition T staging should be considered for treatment escalation to experimental therapies in clinical trials. Conclusion: Nomograms built based on prognostic adverse pathological features can be used within each NCCN risk group to fine-tune treatment decisions for OSCC patients. [ABSTRACT FROM AUTHOR]

Background: Despite revised staging criteria, stratification of patients with advanced oral squamous cell carcinoma (OSCC) remains difficult. Well‐established features like perineural invasion (PNI), differentiation, and lymphovascular‐invasion (LVI) are controversial, and hence omitted from staging. We endeavor to better stratify this cohort by identifying predictors of survival in advanced OSCC (T3‐4). Methods: Seven hundred and forty‐two patients with T3‐4 OSCC underwent surgery from 2006 to 2013. Cox regression was performed to determine predictors of overall survival (OS). Results: OS was adversely impacted by PNI (p = 0.046), LVI (p = 0.038), moderate/poor differentiation (p = 0.001), close/involved surgical margins (p = 0.002), pT (p = 0.034), and pN (p < 0.001). The cumulative number of adverse histopathological features predicted poorer OS; HR 2.64 (CI 1.42–4.90) for one adverse feature and HR 4.23 (CI 2.34–7.67) for ≥2. Conclusion: In advanced OSCC, stratification with histopathologic risk factors can predict survival even in maximally treated patients; adjuvant therapies are unable to entirely mitigate this risk. Incorporation of adverse features into future editions of TNM can improve precision in staging and identify candidates for treatment escalation. [ABSTRACT FROM AUTHOR]

We present for the first time, a fully integrated battery powered RFID integrated circuit (IC) for operation at ultrahigh frequency (UHF) and microwave bands. The battery powered RFID IC can also work as a passive RFID tag without a battery or when the battery has died (i.e., voltage has dropped below 1.3 V); this novel dual passive and battery operation allays one of the major drawbacks of currently available active tags, namely that the tag cannot be used once the battery has died. When powered by a battery, the current consumption is 700 nA at 1.5 V (400 nA if internal signals are not brought out on testpads). This ultra-low-power consumption permits the use of a very small capacity battery of 100 mA.hr for lifetimes exceeding ten years; as a result a battery tag that is very close to a passive tag both in form factor and cost is made possible. The chip is built on a 1-[micro] m digital CMOS process with dual poly layers, EEPROM and Schottky diodes. The RF threshold power at 2.45 GHz is -19 dBm which is the lowest ever reported threshold power for RFID tags and has a range exceeding 3.5 m under FCC unlicensed operation at the 2.4-GHz microwave band. The low threshold is achieved with architectural choices and low-power circuit design techniques. At 915 MHz, based on the experimentally measured tag impedance (92-j837) and the threshold spec of the tag (200 mV), the theoretical minimum range is 24 m. The tag initially is in a 'low-power' mode to conserve power and when issued the appropriate command, it operates in 'full-power' mode. The chip has on-chip voltage regulators, clock and data recovery circuits, EEPROM and a digital state machine that implements the ISO 18000-4 B protocol in the 'full-power' mode. We provide detailed explanation of the clock recovery circuits and the implementation of the binary sort algorithm, which includes a pseudorandom number generator. Other than the antenna board and a battery, no external components are used. Index Terms--18000-4B, active, battery, International Standards Organization (ISO), microwave, passive, RFID, Schottky diode, tag, ultra-high frequency (UHF), ultra-low power.

Purpose: The primary objective of this prospective cohort study was to evaluate primary and secondary implant stability values of implants placed into healed free fibula-reconstructed mandibles. The secondary objective was to investigate whether a correlation existed between primary implant stability and the quantity of cortical bone-implant contact and of cortical fibular bone in contact with the implants. Materials and Methods: The study participants are patients enrolled in a registered trial (No. CTRI/2012/07/002764). Patients with healed segmental mandibular reconstruction by means of free fibula flaps were chosen to undergo implant-supported dental rehabilitation. Preoperatively, computed tomography (CT) scans were used to measure the quality (density) of bone in Hounsfield units (HU). Implant stability quotients (ISQ) were obtained at the time of implant placement (primary), as well as 6 months later (secondary). The amount of cortical bone-to-implant contact (CBIC) was measured from postoperative panoramic radiographs using image analysis software (Image J). A paired t test was used to compare the mean primary and secondary ISQ values. Pearson correlation was used to determine the relationship between primary ISQ and CBIC. Statistical significance was set at the 5% level (P < .05). Results: Of 54 patients (156 implants) who were enrolled, 35 (104 implants) were assessed. The reasons for exclusion were as follows: 11 patients (34 implants) had inadequate preoperative CT documentation, 7 patients (16 implants) had unreadable panoramic radiographs, and 1 patient (2 implants) had partial necrosis of the flap. The mean primary and secondary ISQ values were 78.1 (SD = 4.3) and 77.8 (SD = 4.6), respectively (P = .348). The mean density of cortical fibula bone was 1,344 HU (SD = 93.6). The mean CBIC was 3.7 mm (SD = 1.4). The study results showed a positive correlation between CBIC and implant stability (r = 0.274, P = .005). Conclusion: This study showed no differences in primary and secondary ISQ values of implants placed into the fibula. Implant stability quotients correlated positively with CBIC. [ABSTRACT FROM AUTHOR]

Kurokawa's method of calculating the power reflection coefficient from the Smith chart in the situation when one complex impedance is directly connected to another is applied to passive RFID tag design, where power reflection is important, as it determines the tag characteristics. The performance analysis of a specific RFID tag is presented together with experimental data, which is in close agreement with the theory. Index Terms--Antennas, integrated circuits (ICs), power reflection, RF identification (RFID).