Background
The overall prognosis of oral cancer remains poor because over half of patients are diagnosed at advanced-stages. Previously reported screening and earlier detection methods for oral cancer still largely rely on health workers’ clinical experience and as yet there is no established method. We aimed to develop a rapid, non-invasive, cost-effective, and easy-to-use deep learning approach for identifying oral cavity squamous cell carcinoma (OCSCC) patients using photographic images.
Methods
We developed an automated deep learning algorithm using cascaded convolutional neural networks to detect OCSCC from photographic images. We included all biopsy-proven OCSCC photographs and normal controls of 44,409 clinical images collected from 11 hospitals around China between April 12, 2006, and Nov 25, 2019. We trained the algorithm on a randomly selected part of this dataset (development dataset) and used the rest for testing (internal validation dataset). Additionally, we curated an external validation dataset comprising clinical photographs from six representative journals in the field of dentistry and oral surgery. We also compared the performance of the algorithm with that of seven oral cancer specialists on a clinical validation dataset. We used the pathological reports as gold standard for OCSCC identification. We evaluated the algorithm performance on the internal, external, and clinical validation datasets by calculating the area under the receiver operating characteristic curves (AUCs), accuracy, sensitivity, and specificity with two-sided 95% CIs.
Findings
1469 intraoral photographic images were used to validate our approach. The deep learning algorithm achieved an AUC of 0·983 (95% CI 0·973–0·991), sensitivity of 94·9% (0·915–0·978), and specificity of 88·7% (0·845–0·926) on the internal validation dataset (
n
= 401), and an AUC of 0·935 (0·910–0·957), sensitivity of 89·6% (0·847–0·942) and specificity of 80·6% (0·757–0·853) on the external validation dataset (
n
= 402). For a secondary analysis on the internal validation dataset, the algorithm presented an AUC of 0·995 (0·988–0·999), sensitivity of 97·4% (0·932–1·000) and specificity of 93·5% (0·882–0·979) in detecting early-stage OCSCC. On the clinical validation dataset (
n
= 666), our algorithm achieved comparable performance to that of the average oral cancer expert in terms of accuracy (92·3% [0·902–0·943]
vs
92.4% [0·912–0·936]), sensitivity (91·0% [0·879–0·941]
vs
91·7% [0·898–0·934]), and specificity (93·5% [0·909–0·960]
vs
93·1% [0·914–0·948]). The algorithm also achieved significantly better performance than that of the average medical student (accuracy of 87·0% [0·855–0·885], sensitivity of 83·1% [0·807–0·854], and specificity of 90·7% [0·889–0·924]) and the average non-medical student (accuracy of 77·2% [0...
We consider the problem of expanding arc capacities in a network subject to demand and travel time uncertainty. We propose a robust optimization approach to obtain capacity expansion solutions that are insensitive to this uncertainty. Our results show that, under reasonable assumptions for network flow applications, such robust solutions can be computed by solving tractable conic linear problems. For example, the robust solution for a multicommodity flow problem is obtained by solving a linear program if the problem has a single source and sink per commodity and the uncertainty in demand and travel time is given by independent bounded polyhedral sets. Preliminary computational results show that the robust solution is attractive, as it can reduce the worst case cost by more than 20%, while incurring a 5% loss in optimality when compared to the optimal solution of a representative scenario.
In order to improve service reliability, many transit agencies add a significant amount of slack in the schedule. However, too much slack in the schedule reduces service frequency. We study the problem of determining the optimal slack that minimizes the passengers' expected waiting times under schedule-based control. By associating with a D/G/c queue model, we show that the system is stable if slack is added in the schedule. For a single-bus loop transit network, we derive convexity of mean and variance of bus delays and provide an exact solution if the travel time is exponentially distributed. For the case of multiple buses and other travel time distributions, we provide several approximation approaches and compare them to simulation results. The simulation results show that our approximations are good for interval of appropriate slack, which often contains the optimal value.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.