An artificial intelligence‐driven agent for real‐time head‐and‐neck IMRT plan generation using conditional generative adversarial network (cGAN). Issue 6 (25th April 2021)

• Record Type: Journal Article
• Authors: Li, Xinyi; Wang, Chunhao; Sheng, Yang; Zhang, Jiahan; Wang, Wentao; Yin, Fang‐Fang; Wu, Qiuwen; Wu, Q. Jackie; Ge, Yaorong
• Is Part Of: Medical physics. Volume 48:Issue 6(2021)
• Page Start: 2714
• Page End: 2723
• Publication Date: 2021-04-25
• Subjects: artificial intelligence -- conditional generative adversarial network -- deep learning -- fluence map prediction -- head‐and‐neck cancer -- intensity‐modulated radiation therapy -- treatment planning; Medical physics -- Periodicals; Medical physics; Geneeskunde; Natuurkunde; Toepassingen; Biophysics; Periodicals; Periodicals; Electronic journals; 610.153
• DOI: 10.1002/mp.14770
• Languages: English
• ISSNs: 0094-2405

Abstract

Abstract : Purpose: To develop an artificial intelligence (AI) agent for fully automated rapid head‐and‐neck intensity‐modulated radiation therapy (IMRT) plan generation without time‐consuming dose‐volume‐based inverse planning. Methods: This AI agent was trained via implementing a conditional generative adversarial network (cGAN) architecture. The generator, PyraNet, is a novel deep learning network that implements 28 classic ResNet blocks in pyramid‐like concatenations. The discriminator is a customized four‐layer DenseNet. The AI agent first generates multiple customized two‐dimensional projections at nine template beam angles from a patient's three‐dimensional computed tomography (CT) volume and structures. These projections are then stacked as four‐dimensional inputs of PyraNet, from which nine radiation fluence maps of the corresponding template beam angles are generated simultaneously. Finally, the predicted fluence maps are automatically postprocessed by Gaussian deconvolution operations and imported into a commercial treatment planning system (TPS) for plan integrity check and visualization. The AI agent was built and tested upon 231 oropharyngeal IMRT plans from a TPS plan library. 200/16/15 plans were assigned for training/validation/testing, respectively. Only the primary plans in the sequential boost regime were studied. All plans were normalized to 44 Gy prescription (2 Gy/fx). A customized Harr wavelet loss was adopted for fluence map comparison during the training of the PyraNet. For test cases, isodose distributions in AI plans and TPS plans were qualitatively evaluated for overall dose distributions. Key dosimetric metrics were compared by Wilcoxon signed‐rank tests with a significance level of 0.05. Results: All 15 AI plans were successfully generated. Isodose gradients outside of PTV in AI plans were comparable to those of the TPS plans. After PTV coverage normalization, Dmean of left parotid (DAI = 23.1 ± 2.4 Gy; DTPS = 23.1 ± 2.0 Gy), right parotid (DAI = 23.8 ± 3.0 Gy; DTPS = 23.9 ± 2.3 Gy), and oral cavity (DAI = 24.7 ± 6.0 Gy; DTPS = 23.9 ± 4.3 Gy) in the AI plans and the TPS plans were comparable without statistical significance. AI plans achieved comparable results for maximum dose at 0.01cc of brainstem (DAI = 15.0 ± 2.1 Gy; DTPS = 15.5 ± 2.7 Gy) and cord + 5mm (DAI = 27.5 ± 2.3 Gy; DTPS = 25.8 ± 1.9 Gy) without clinically relevant differences, but body Dmax results (DAI = 121.1 ± 3.9 Gy; DTPS = 109.0 ± 0.9 Gy) were higher than the TPS plan results. The AI agent needed ~3 s for predicting fluence maps of an IMRT plan. Conclusions: With rapid and fully automated execution, the developed AI agent can generate complex head‐and‐neck IMRT plans with acceptable dosimetry quality. This approach holds great potential for clinical applications in preplanning decision‐making and real‐time planning.