Cancer Informatics for Cancer Centers: Sharing Ideas on How to Build an Artificial Intelligence-Ready Informatics Ecosystem for Radiation Oncology.

In August 2022, the Cancer Informatics for Cancer Centers brought together cancer informatics leaders for its biannual symposium, Precision Medicine Applications in Radiation Oncology, co-chaired by Quynh-Thu Le, MD (Stanford University), and Walter J. Curran, MD (GenesisCare). Over the course of 3 days, presenters discussed a range of topics relevant to radiation oncology and the cancer informatics community more broadly, including biomarker development, decision support algorithms, novel imaging tools, theranostics, and artificial intelligence (AI) for the radiotherapy workflow. Since the symposium, there has been an impressive shift in the promise and potential for integration of AI in clinical care, accelerated in large part by major advances in generative AI. AI is now poised more than ever to revolutionize cancer care. Radiation oncology is a field that uses and generates a large amount of digital data and is therefore likely to be one of the first fields to be transformed by AI. As experts in the collection, management, and analysis of these data, the informatics community will take a leading role in ensuring that radiation oncology is prepared to take full advantage of these technological advances. In this report, we provide highlights from the symposium, which took place in Santa Barbara, California, from August 29 to 31, 2022. We discuss lessons learned from the symposium for data acquisition, management, representation, and sharing, and put these themes into context to prepare radiation oncology for the successful and safe integration of AI and informatics technologies.

Radiotherapy Plus Cisplatin With or Without Lapatinib for Non-Human Papillomavirus Head and Neck Carcinoma: A Phase 2 Randomized Clinical Trial.

Importance: Patients with locally advanced non-human papillomavirus (HPV) head and neck cancer (HNC) carry an unfavorable prognosis. Chemoradiotherapy (CRT) with cisplatin or anti-epidermal growth factor receptor (EGFR) antibody improves overall survival (OS) of patients with stage III to IV HNC, and preclinical data suggest that a small-molecule tyrosine kinase inhibitor dual EGFR and ERBB2 (formerly HER2 or HER2/neu) inhibitor may be more effective than anti-EGFR antibody therapy in HNC. Objective: To examine whether adding lapatinib, a dual EGFR and HER2 inhibitor, to radiation plus cisplatin for frontline therapy of stage III to IV non-HPV HNC improves progression-free survival (PFS). Design, Setting, and Participants: This multicenter, phase 2, double-blind, placebo-controlled randomized clinical trial enrolled 142 patients with stage III to IV carcinoma of the oropharynx (p16 negative), larynx, and hypopharynx with a Zubrod performance status of 0 to 1 who met predefined blood chemistry criteria from October 18, 2012, to April 18, 2017 (median follow-up, 4.1 years). Data analysis was performed from December 1, 2020, to December 4, 2020. Intervention: Patients were randomized (1:1) to 70 Gy (6 weeks) plus 2 cycles of cisplatin (every 3 weeks) plus either 1500 mg per day of lapatinib (CRT plus lapatinib) or placebo (CRT plus placebo). Main Outcomes and Measures: The primary end point was PFS, with 69 events required. Progression-free survival rates between arms for all randomized patients were compared by 1-sided log-rank test. Secondary end points included OS. Results: Of the 142 patients enrolled, 127 (median [IQR] age, 58 [53-63] years; 98 [77.2%] male) were randomized; 63 to CRT plus lapatinib and 64 to CRT plus placebo. Final analysis did not suggest improvement in PFS (hazard ratio, 0.91; 95% CI, 0.56-1.46; P = .34) or OS (hazard ratio, 1.06; 95% CI, 0.61-1.86; P = .58) with the addition of lapatinib. There were no significant differences in grade 3 to 4 acute adverse event rates (83.3% [95% CI, 73.9%-92.8%] with CRT plus lapatinib vs 79.7% [95% CI, 69.4%-89.9%] with CRT plus placebo; P = .64) or late adverse event rates (44.4% [95% CI, 30.2%-57.8%] with CRT plus lapatinib vs 40.8% [95% CI, 27.1%-54.6%] with CRT plus placebo; P = .84). Conclusion and Relevance: In this randomized clinical trial, dual EGFR-ERBB2 inhibition with lapatinib did not appear to enhance the benefit of CRT. Although the results of this trial indicate that accrual to a non-HPV HNC-specific trial is feasible, new strategies must be investigated to improve the outcome for this population with a poor prognosis. Trial Registration: ClinicalTrials.gov Identifier: NCT01711658.

Pathologic Complete Response and Clinical Outcomes in Patients With Localized Soft Tissue Sarcoma Treated With Neoadjuvant Chemoradiotherapy or Radiotherapy: The NRG/RTOG 9514 and 0630 Nonrandomized Clinical Trials.

Importance: Pathologic complete response (pCR) may be associated with prognosis in patients with soft tissue sarcoma (STS). Objective: We sought to determine the prognostic significance of pCR on survival outcomes in STS for patients receiving neoadjuvant chemoradiotherapy (CT-RT) (Radiation Therapy Oncology Group [RTOG] 9514) or preoperative image-guided radiotherapy alone (RT, RTOG 0630) and provide a long-term update of RTOG 0630. Design, Setting, and Participants: RTOG has completed 2 multi-institutional, nonrandomized phase 2 clinical trials for patients with localized STS. One hundred forty-three eligible patients from RTOG 0630 (n = 79) and RTOG 9514 (n = 64) were included in this ancillary analysis of pCR and 79 patients from RTOG 0630 were evaluated for long-term outcomes. Intervention: Patients in trial 9514 received CT interdigitated with RT, whereas those in trial 0630 received preoperative RT alone. Main Outcomes and Measures: Overall and disease-free survival (OS and DFS) rates were estimated by the Kaplan-Meier method. Hazard ratios (HRs) and P values were estimated by multivariable Cox model stratified by study, where possible; otherwise, P values were calculated by stratified log-rank test. Analysis took place between December 14, 2016, to April 13, 2017. Results: Overall there were 42 (53.2%) men; 68 (86.1%) were white; with a mean (SD) age of 59.6 (14.5) years. For RTOG 0630, at median follow-up of 6.0 years, there was 1 new in-field recurrence and 1 new distant failure since the initial report. From both studies, 123 patients were evaluable for pCR: 14 of 51 (27.5%) in trial 9514 and 14 of 72 (19.4%) in trial 0630 had pCR. Five-year OS was 100% for patients with pCR vs 76.5% (95% CI, 62.3%-90.8%) and 56.4% (95% CI, 43.3%-69.5%) for patients with less than pCR in trials 9514 and 0630, respectively. Overall, pCR was associated with improved OS (P = .01) and DFS (HR, 4.91; 95% CI, 1.51-15.93; P = .008) relative to less than pCR. Five-year local failure rate was 0% in patients with pCR vs 11.7% (95% CI, 3.6%-25.1%) and 9.1% (95% CI, 3.3%-18.5%) for patients with less than pCR in 9514 and 0630, respectively. Histologic types other than leiomyosarcoma, liposarcoma, and myxofibrosarcoma were associated with worse OS (HR, 2.24; 95% CI, 1.12-4.45). Conclusions and Relevance: This ancillary analysis of 2 nonrandomized clinical trials found that pCR was associated with improved survival in patients with STS and should be considered as a prognostic factor of clinical outcomes for future studies. Trial Registration: ClinicalTrials.gov Identifiers: RTOG 0630 (NCT00589121); RTOG 9514 (NCT00002791).

Population, Clinical, and Scientific Impact of National Cancer Institute's National Clinical Trials Network Treatment Studies.

PURPOSE: In the United States, the National Cancer Institute National Cancer Clinical Trials Network (NCTN) groups have conducted publicly funded oncology research for 50 years. The combined impact of all adult network group trials has never been systematically examined. METHODS: We identified randomized, phase III trials from the adult NCTN groups, reported from 1980 onward, with statistically significant findings for ≥ 1 clinical, time-dependent outcomes. In the subset of trials in which the experimental arm improved overall survival, gains in population life-years were estimated by deriving trial-specific hazard functions and hazard ratios to estimate the experimental treatment benefit and then mapping this trial-level benefit onto the US cancer population using registry and life-table data. Scientific impact was based on citation data from Google Scholar. Federal investment costs per life-year gained were estimated. The results were derived through December 31, 2020. RESULTS: One hundred sixty-two trials comprised of 108,334 patients were analyzed, representing 29.8% (162/544) of trials conducted. The most common cancers included breast (34), gynecologic (28), and lung (14). The trials were cited 165,336 times (mean, 62.2 citations/trial/year); 87.7% of trials were cited in cancer care guidelines in favor of the recommended treatment. These studies were estimated to have generated 14.2 million (95% CI, 11.5 to 16.5 million) additional life-years to patients with cancer, with projected gains of 24.1 million (95% CI, 19.7 to 28.2 million) life-years by 2030. The federal investment cost per life-year gained through 2020 was $326 in US dollars. CONCLUSION: NCTN randomized trials have been widely cited and are routinely included in clinical guidelines. Moreover, their conduct has predicted substantial improvements in overall survival in the United States for patients with oncologic disease, suggesting they have contributed meaningfully to this nation's health. These findings demonstrate the critical role of government-sponsored research in extending the lives of patients with cancer.

Depatuxizumab mafodotin in EGFR-amplified newly diagnosed glioblastoma: A phase III randomized clinical trial.

BACKGROUND: Approximately 50% of newly diagnosed glioblastomas (GBMs) harbor epidermal growth factor receptor gene amplification (EGFR-amp). Preclinical and early-phase clinical data suggested efficacy of depatuxizumab mafodotin (depatux-m), an antibody-drug conjugate comprised of a monoclonal antibody that binds activated EGFR (overexpressed wild-type and EGFRvIII-mutant) linked to a microtubule-inhibitor toxin in EGFR-amp GBMs. METHODS: In this phase III trial, adults with centrally confirmed, EGFR-amp newly diagnosed GBM were randomized 1:1 to radiotherapy, temozolomide, and depatux-m/placebo. Corneal epitheliopathy was treated with a combination of protocol-specified prophylactic and supportive measures. There was 85% power to detect a hazard ratio (HR) ≤0.75 for overall survival (OS) at a 2.5% 1-sided significance level (ie traditional two-sided p ≤ 0.05) by log-rank testing. RESULTS: There were 639 randomized patients (median age 60, range 22-84; 62% men). Prespecified interim analysis found no improvement in OS for depatux-m over placebo (median 18.9 vs. 18.7 months, HR 1.02, 95% CI 0.82-1.26, 1-sided p = 0.63). Progression-free survival was longer for depatux-m than placebo (median 8.0 vs. 6.3 months; HR 0.84, 95% confidence interval [CI] 0.70-1.01, p = 0.029), particularly among those with EGFRvIII-mutant (median 8.3 vs. 5.9 months, HR 0.72, 95% CI 0.56-0.93, 1-sided p = 0.002) or MGMT unmethylated (HR 0.77, 95% CI 0.61-0.97; 1-sided p = 0.012) tumors but without an OS improvement. Corneal epitheliopathy occurred in 94% of depatux-m-treated patients (61% grade 3-4), causing 12% to discontinue. CONCLUSIONS: Interim analysis demonstrated no OS benefit for depatux-m in treating EGFR-amp newly diagnosed GBM. No new important safety risks were identified.

Safety of Nivolumab Added to Chemoradiation Therapy Platforms for Intermediate and High-Risk Locoregionally Advanced Head and Neck Squamous Cell Carcinoma: RTOG Foundation 3504.

PURPOSE: Programmed death-1 immune checkpoint blockade improves survival of patients with recurrent/metastatic head and neck squamous cell carcinoma (HNSCC), but the benefits of addition to (chemo)radiation for newly diagnosed patients with HNSCC remain unknown. METHODS AND MATERIALS: We evaluated the safety of nivolumab concomitant with 70 Gy intensity modulated radiation therapy and weekly cisplatin (arm 1), every 3-week cisplatin (arm 2), cetuximab (arm 3), or alone for platinum-ineligible patients (arm 4) in newly diagnosed intermediate- or high-risk locoregionally advanced HNSCC. Patients received nivolumab from 2 weeks prior to radiation therapy until 3 months post-radiation therapy. The primary endpoint was dose-limiting toxicity (DLT). If ≤2 of the first 8 evaluable patients experienced a DLT, an arm was considered safe. Secondary endpoints included toxicity and feasibility of adjuvant nivolumab to 1 year, defined as all 7 additional doses received by ≥4 of the first 8 evaluable patients across arms. RESULTS: Of 39 patients (10 in arms 1, 3, 4 and 9 in arm 2), 72% had T3-4 tumors, 85% had N2-3 nodal disease, and 67% had >10 pack-years of smoking. There were no DLTs in arms 1 and 2, 1 in arm 3 (mucositis), and 2 in arm 4 (lipase elevation and mucositis in 1 and fatigue in another). The most common grade ≥3 nivolumab-related adverse events were lipase increase, mucositis, diarrhea, lymphopenia, hyponatremia, leukopenia, fatigue, and serum amylase increase. Adjuvant nivolumab was feasible as defined in the protocol. CONCLUSIONS: Concomitant nivolumab with the 4 tested regimens was safe for patients with intermediate- and high-risk HNSCC, and subsequent adjuvant nivolumab was feasible as defined (NCT02764593).

Preliminary efficacy and safety analysis: 12-month results in 83 patients using a novel approach of widefield radiation therapy for extensive skin field cancerization with or without keratinocyte cancers.

PURPOSE: Evaluate the use of widefield radiation therapy (RT) in the management of extensive skin field cancerization (ESFC) with/without keratinocyte cancer (KC). METHODS: The National Dermatology Radiation Oncology Registry is a multidisciplinary collaboration (dermatologists and radiation oncologists). It captures disease description, prior therapies, radiation prescription, clinical effect, skin cosmesis scores, and toxicity data. This analysis included 12-month follow-up data on 89 treated fields from a subset of 83 patients. RESULTS: Clinical success (>90% field clearance) was 96% (ESFC, n = 63) and 88% (ESFC with KC, n = 26). Complete lesion response was seen in 96% of evaluable (n = 25) ESFC with KC. Recurrence (4/89 [5%]) and appearance of new lesions (10/89 [11%]) were minimal. Cosmetic outcome was excellent/good in 98% ESFC and 96% ESFC with KC. Grade 1-2 acute radiation dermatitis occurred in up to 80% of treated fields. The frequency of Grade 3 acute skin toxicities was low. CONCLUSIONS: Registry data demonstrate the potential for widefield RT to treat patients with significant skin pathology who have exhausted other therapies and require durable, minimally invasive treatment options. At 12 months, observed clinical success rates were higher than those reported for topical interventions for ESFC. Ongoing follow-up is required to determine longer term outcomes.

Learning-based synthetic dual energy CT imaging from single energy CT for stopping power ratio calculation in proton radiation therapy.

OBJECTIVE: Dual energy CT (DECT) has been shown to estimate stopping power ratio (SPR) map with a higher accuracy than conventional single energy CT (SECT) by obtaining the energy dependence of photon interactions. This work presents a learning-based method to synthesize DECT images from SECT image for proton radiotherapy. METHODS: The proposed method uses a residual attention generative adversarial network. Residual blocks with attention gates were used to force the model to focus on the difference between DECT images and SECT images. To evaluate the accuracy of the method, we retrospectively investigated 70 head-and-neck cancer patients whose DECT and SECT scans were acquired simultaneously. The model was trained to generate both a high and low energy DECT image based on a SECT image. The generated synthetic low and high DECT images were evaluated against the true DECT images using leave-one-out cross-validation. To evaluate our method in the context of a practical application, we generated SPR maps from synthetic DECT (sDECT) using a dual-energy based stoichiometric method and compared the SPR maps to those generated from DECT. A dosimetric comparison for dose obtained from DECT was performed against that derived from sDECT. RESULTS: The mean of mean absolute error, peak signal-to-noise ratio and normalized cross-correlation for the synthetic high and low energy CT images was 36.9 HU, 29.3 dB, 0.96 and 35.8 HU, 29.2 dB, and 0.96, respectively. The corresponding SPR maps generated from synthetic DECT showed an average normalized mean square deviation of about 1% with reduced noise level and artifacts than those from original DECT. Dose-volume histogram (DVH) metrics for the clinical target volume agree within 1% between the DECT and sDECT calculated dose. CONCLUSION: Our method synthesized accurate DECT images and showed a potential feasibility for proton SPR map generation. ADVANCES IN KNOWLEDGE: This study investigated a learning-based method to synthesize DECT images from SECT image for proton radiotherapy.

Artificial Intelligence in Quantitative Ultrasound Imaging: A Survey.

Quantitative ultrasound (QUS) imaging is a safe, reliable, inexpensive, and real-time technique to extract physically descriptive parameters for assessing pathologies. Compared with other major imaging modalities such as computed tomography and magnetic resonance imaging, QUS suffers from several major drawbacks: poor image quality and inter- and intra-observer variability. Therefore, there is a great need to develop automated methods to improve the image quality of QUS. In recent years, there has been increasing interest in artificial intelligence (AI) applications in medical imaging, and a large number of research studies in AI in QUS have been conducted. The purpose of this review is to describe and categorize recent research into AI applications in QUS. We first introduce the AI workflow and then discuss the various AI applications in QUS. Finally, challenges and future potential AI applications in QUS are discussed.

Surgical Outcomes for Early Stage Non-small Cell Lung Cancer at Facilities With Stereotactic Body Radiation Therapy Programs.

BACKGROUND: Patients undergoing surgery for early stage non-small cell lung cancer (NSCLC) may be at high risk for postoperative mortality. Access to stereotactic body radiation therapy (SBRT) may facilitate more appropriate patient selection for surgery. RESEARCH QUESTION: Is postoperative mortality associated with early stage NSCLC lower at facilities with higher use of SBRT? STUDY DESIGN AND METHODS: Patients with early stage NSCLC reported to the National Cancer Database between 2004 and 2015 were included. Use of SBRT was defined by each facility's SBRT experience (in years) and SBRT to surgery volume ratios. Multivariate logistic regression was used to test for the associations between SBRT use and postoperative mortality. RESULTS: The study cohort consisted of 202,542 patients who underwent surgical resection of cT1-T2N0M0 NSCLC tumors. The 90-day postoperative mortality rate declined during the study period from 4.6% to 2.6% (P < .001), the proportion of facilities that used SBRT increased from 4.6% to 77.5% (P < .001), and the proportion of patients treated with SBRT increased from 0.7% to 15.4% (P < .001). On multivariate analysis, lower 90-day postoperative mortality rates were observed at facilities with > 6 years of SBRT experience (OR, 0.84; 95% CI, 0.76-0.94; P = .003) and SBRT to surgery volume ratios of more than 17% (OR, 0.85; 95% CI, 0.79-0.92; P < .001). Ninety-day mortality also was associated with surgical volume, region, year, age, sex, and race, among other covariates. Interaction testing between these covariates showed negative results. INTERPRETATION: Patients who underwent resection for early stage NSCLC at facilities with higher SBRT use showed lower rates of postoperative mortality. These findings suggest that the availability and use of SBRT may improve the selection of patients for surgery who are predicted to be at high risk of postoperative mortality.

Multi-organ auto-delineation in head-and-neck MRI for radiation therapy using regional convolutional neural network.

Magnetic resonance imaging (MRI) allows accurate and reliable organ delineation for many disease sites in radiation therapy because MRI is able to offer superb soft-tissue contrast. Manual organ-at-risk delineation is labor-intensive and time-consuming. This study aims to develop a deep-learning-based automated multi-organ segmentation method to release the labor and accelerate the treatment planning process for head-and-neck (HN) cancer radiotherapy. A novel regional convolutional neural network (R-CNN) architecture, namely, mask scoring R-CNN, has been developed in this study. In the proposed model, a deep attention feature pyramid network is used as a backbone to extract the coarse features given by MRI, followed by feature refinement using R-CNN. The final segmentation is obtained through mask and mask scoring networks taking those refined feature maps as input. With the mask scoring mechanism incorporated into conventional mask supervision, the classification error can be highly minimized in conventional mask R-CNN architecture. A cohort of 60 HN cancer patients receiving external beam radiation therapy was used for experimental validation. Five-fold cross-validation was performed for the assessment of our proposed method. The Dice similarity coefficients of brain stem, left/right cochlea, left/right eye, larynx, left/right lens, mandible, optic chiasm, left/right optic nerve, oral cavity, left/right parotid, pharynx, and spinal cord were 0.89 ± 0.06, 0.68 ± 0.14/0.68 ± 0.18, 0.89 ± 0.07/0.89 ± 0.05, 0.90 ± 0.07, 0.67 ± 0.18/0.67 ± 0.10, 0.82 ± 0.10, 0.61 ± 0.14, 0.67 ± 0.11/0.68 ± 0.11, 0.92 ± 0.07, 0.85 ± 0.06/0.86 ± 0.05, 0.80 ± 0.13, and 0.77 ± 0.15, respectively. After the model training, all OARs can be segmented within 1 min.

MRI classification using semantic random forest with auto-context model.

BACKGROUND: It is challenging to differentiate air and bone on MR images of conventional sequences due to their low contrast. We propose to combine semantic feature extraction under auto-context manner into random forest to improve reasonability of the MRI segmentation for MRI-based radiotherapy treatment planning or PET attention correction. METHODS: We applied a semantic classification random forest (SCRF) method which consists of a training stage and a segmentation stage. In the training stage, patch-based MRI features were extracted from registered MRI-CT training images, and the most informative elements were selected via feature selection to train an initial random forest. The rest sequence of random forests was trained by a combination of MRI feature and semantic feature under an auto-context manner. During segmentation, the MRI patches were first fed into these random forests to derive patch-based segmentation. By using patch fusion, the final end-to-end segmentation was obtained. RESULTS: The Dice similarity coefficient (DSC) for air, bone and soft tissue classes obtained via proposed method were 0.976±0.007, 0.819±0.050 and 0.932±0.031, compared to 0.916±0.099, 0.673±0.151 and 0.830±0.083 with random forest (RF), and 0.942±0.086, 0.791±0.046 and 0.917±0.033 with U-Net. SCRF also outperformed the competing methods in sensitivity and specificity for all three structure types. CONCLUSIONS: The proposed method accurately segmented bone, air and soft tissue. It is promising in facilitating advanced MR application in diagnosis and therapy.

Deep learning-based motion tracking using ultrasound images.

PURPOSE: Ultrasound (US) imaging is an established imaging modality capable of offering video-rate volumetric images without ionizing radiation. It has the potential for intra-fraction motion tracking in radiation therapy. In this study, a deep learning-based method has been developed to tackle the challenges in motion tracking using US imaging. METHODS: We present a Markov-like network, which is implemented via generative adversarial networks, to extract features from sequential US frames (one tracked frame followed by untracked frames) and thereby estimate a set of deformation vector fields (DVFs) through the registration of the tracked frame and the untracked frames. The positions of the landmarks in the untracked frames are finally determined by shifting landmarks in the tracked frame according to the estimated DVFs. The performance of the proposed method was evaluated on the testing dataset by calculating the tracking error (TE) between the predicted and ground truth landmarks on each frame. RESULTS: The proposed method was evaluated using the MICCAI CLUST 2015 dataset which was collected using seven US scanners with eight types of transducers and the Cardiac Acquisitions for Multi-structure Ultrasound Segmentation (CAMUS) dataset which was acquired using GE Vivid E95 ultrasound scanners. The CLUST dataset contains 63 2D and 22 3D US image sequences respectively from 42 and 18 subjects, and the CAMUS dataset includes 2D US images from 450 patients. On CLUST dataset, our proposed method achieved a mean tracking error of 0.70 ± 0.38 mm for the 2D sequences and 1.71 ± 0.84 mm for the 3D sequences for those public available annotations. And on CAMUS dataset, a mean tracking error of 0.54 ± 1.24 mm for the landmarks in the left atrium was achieved. CONCLUSIONS: A novel motion tracking algorithm using US images based on modern deep learning techniques has been demonstrated in this study. The proposed method can offer millimeter-level tumor motion prediction in real time, which has the potential to be adopted into routine tumor motion management in radiation therapy.

Synthetic CT-aided multiorgan segmentation for CBCT-guided adaptive pancreatic radiotherapy.

PURPOSE: The delineation of organs at risk (OARs) is fundamental to cone-beam CT (CBCT)-based adaptive radiotherapy treatment planning, but is time consuming, labor intensive, and subject to interoperator variability. We investigated a deep learning-based rapid multiorgan delineation method for use in CBCT-guided adaptive pancreatic radiotherapy. METHODS: To improve the accuracy of OAR delineation, two innovative solutions have been proposed in this study. First, instead of directly segmenting organs on CBCT images, a pretrained cycle-consistent generative adversarial network (cycleGAN) was applied to generating synthetic CT images given CBCT images. Second, an advanced deep learning model called mask-scoring regional convolutional neural network (MS R-CNN) was applied on those synthetic CT to detect the positions and shapes of multiple organs simultaneously for final segmentation. The OAR contours delineated by the proposed method were validated and compared with expert-drawn contours for geometric agreement using the Dice similarity coefficient (DSC), 95th percentile Hausdorff distance (HD95), mean surface distance (MSD), and residual mean square distance (RMS). RESULTS: Across eight abdominal OARs including duodenum, large bowel, small bowel, left and right kidneys, liver, spinal cord, and stomach, the geometric comparisons between automated and expert contours are as follows: 0.92 (0.89-0.97) mean DSC, 2.90 mm (1.63-4.19 mm) mean HD95, 0.89 mm (0.61-1.36 mm) mean MSD, and 1.43 mm (0.90-2.10 mm) mean RMS. Compared to the competing methods, our proposed method had significant improvements (p < 0.05) in all the metrics for all the eight organs. Once the model was trained, the contours of eight OARs can be obtained on the order of seconds. CONCLUSIONS: We demonstrated the feasibility of a synthetic CT-aided deep learning framework for automated delineation of multiple OARs on CBCT. The proposed method could be implemented in the setting of pancreatic adaptive radiotherapy to rapidly contour OARs with high accuracy.

Deep learning-based thoracic CBCT correction with histogram matching.

Kilovoltage cone-beam computed tomography (CBCT)-based image-guided radiation therapy (IGRT) is used for daily delivery of radiation therapy, especially for stereotactic body radiation therapy (SBRT), which imposes particularly high demands for setup accuracy. The clinical applications of CBCTs are constrained, however, by poor soft tissue contrast, image artifacts, and instability of Hounsfield unit (HU) values. Here, we propose a new deep learning-based method to generate synthetic CTs (sCT) from thoracic CBCTs. A deep-learning model which integrates histogram matching (HM) into a cycle-consistent adversarial network (Cycle-GAN) framework, called HM-Cycle-GAN, was trained to learn mapping between thoracic CBCTs and paired planning CTs. Perceptual supervision was adopted to minimize blurring of tissue interfaces. An informative maximizing loss was calculated by feeding CBCT into the HM-Cycle-GAN to evaluate the image histogram matching between the planning CTs and the sCTs. The proposed algorithm was evaluated using data from 20 SBRT patients who each received 5 fractions and therefore 5 thoracic CBCTs. To reduce the effect of anatomy mismatch, original CBCT images were pre-processed via deformable image registrations with the planning CT before being used in model training and result assessment. We used planning CTs as ground truth for the derived sCTs from the correspondent co-registered CBCTs. The mean absolute error (MAE), peak signal-to-noise ratio (PSNR), and normalized cross-correlation (NCC) indices were adapted as evaluation metrics of the proposed algorithm. Assessments were done using Cycle-GAN as the benchmark. The average MAE, PSNR, and NCC of the sCTs generated by our method were 66.2 HU, 30.3 dB, and 0.95, respectively, over all CBCT fractions. Superior image quality and reduced noise and artifact severity were seen using the proposed method compared to the results from the standard Cycle-GAN method. Our method could therefore improve the accuracy of IGRT and corrected CBCTs could help improve online adaptive RT by offering better contouring accuracy and dose calculation.

Automated delineation of head and neck organs at risk using synthetic MRI-aided mask scoring regional convolutional neural network.

PURPOSE: Auto-segmentation algorithms offer a potential solution to eliminate the labor-intensive, time-consuming, and observer-dependent manual delineation of organs-at-risk (OARs) in radiotherapy treatment planning. This study aimed to develop a deep learning-based automated OAR delineation method to tackle the current challenges remaining in achieving reliable expert performance with the state-of-the-art auto-delineation algorithms. METHODS: The accuracy of OAR delineation is expected to be improved by utilizing the complementary contrasts provided by computed tomography (CT) (bony-structure contrast) and magnetic resonance imaging (MRI) (soft-tissue contrast). Given CT images, synthetic MR images were firstly generated by a pre-trained cycle-consistent generative adversarial network. The features of CT and synthetic MRI were then extracted and combined for the final delineation of organs using mask scoring regional convolutional neural network. Both in-house and public datasets containing CT scans from head-and-neck (HN) cancer patients were adopted to quantitatively evaluate the performance of the proposed method against current state-of-the-art algorithms in metrics including Dice similarity coefficient (DSC), 95th percentile Hausdorff distance (HD95), mean surface distance (MSD), and residual mean square distance (RMS). RESULTS: Across all of 18 OARs in our in-house dataset, the proposed method achieved an average DSC, HD95, MSD, and RMS of 0.77 (0.58-0.90), 2.90 mm (1.32-7.63 mm), 0.89 mm (0.42-1.85 mm), and 1.44 mm (0.71-3.15 mm), respectively, outperforming the current state-of-the-art algorithms by 6%, 16%, 25%, and 36%, respectively. On public datasets, for all nine OARs, an average DSC of 0.86 (0.73-0.97) were achieved, 6% better than the competing methods. CONCLUSION: We demonstrated the feasibility of a synthetic MRI-aided deep learning framework for automated delineation of OARs in HN radiotherapy treatment planning. The proposed method could be adopted into routine HN cancer radiotherapy treatment planning to rapidly contour OARs with high accuracy.

Knowledge-based radiation treatment planning: A data-driven method survey.

This paper surveys the data-driven dose prediction methods investigated for knowledge-based planning (KBP) in the last decade. These methods were classified into two major categories-traditional KBP methods and deep-learning (DL) methods-according to their techniques of utilizing previous knowledge. Traditional KBP methods include studies that require geometric or anatomical features to either find the best-matched case(s) from a repository of prior treatment plans or to build dose prediction models. DL methods include studies that train neural networks to make dose predictions. A comprehensive review of each category is presented, highlighting key features, methods, and their advancements over the years. We separated the cited works according to the framework and cancer site in each category. Finally, we briefly discuss the performance of both traditional KBP methods and DL methods, then discuss future trends of both data-driven KBP methods to dose prediction.

Artificial intelligence in tumor subregion analysis based on medical imaging: A review.

Medical imaging is widely used in the diagnosis and treatment of cancer, and artificial intelligence (AI) has achieved tremendous success in medical image analysis. This paper reviews AI-based tumor subregion analysis in medical imaging. We summarize the latest AI-based methods for tumor subregion analysis and their applications. Specifically, we categorize the AI-based methods by training strategy: supervised and unsupervised. A detailed review of each category is presented, highlighting important contributions and achievements. Specific challenges and potential applications of AI in tumor subregion analysis are discussed.

Fully automated segmentation of brain tumor from multiparametric MRI using 3D context deep supervised U-Net.

PURPOSE: Owing to histologic complexities of brain tumors, its diagnosis requires the use of multimodalities to obtain valuable structural information so that brain tumor subregions can be properly delineated. In current clinical workflow, physicians typically perform slice-by-slice delineation of brain tumor subregions, which is a time-consuming process and also more susceptible to intra- and inter-rater variabilities possibly leading to misclassification. To deal with this issue, this study aims to develop an automatic segmentation of brain tumor in MR images using deep learning. METHOD: In this study, we develop a context deep-supervised U-Net to segment brain tumor subregions. A context block which aggregates multiscale contextual information for dense segmentation was proposed. This approach enlarges the effective receptive field of convolutional neural networks, which, in turn, improves the segmentation accuracy of brain tumor subregions. We performed the fivefold cross-validation on the Brain Tumor Segmentation Challenge (BraTS) 2020 training dataset. The BraTS 2020 testing datasets were obtained via BraTS online website as a hold-out test. For BraTS, the evaluation system divides the tumor into three regions: whole tumor (WT), tumor core (TC), and enhancing tumor (ET). The performance of our proposed method was compared against two state-of-the-arts CNN networks in terms of segmentation accuracy via Dice similarity coefficient (DSC) and Hausdorff distance (HD). The tumor volumes generated by our proposed method were compared with manually contoured volumes via Bland-Altman plots and Pearson analysis. RESULTS: The proposed method achieved the segmentation results with a DSC of 0.923 ± 0.047, 0.893 ± 0.176, and 0.846 ± 0.165 and a 95% HD95 of 3.946 ± 7.041, 3.981 ± 6.670, and 10.128 ± 51.136 mm on WT, TC, and ET, respectively. Experimental results demonstrate that our method achieved comparable to significantly (p < 0.05) better segmentation accuracies than other two state-of-the-arts CNN networks. Pearson correlation analysis showed a high positive correlation between the tumor volumes generated by proposed method and manual contour. CONCLUSION: Overall qualitative and quantitative results of this work demonstrate the potential of translating proposed technique into clinical practice for segmenting brain tumor subregions, and further facilitate brain tumor radiotherapy workflow.

Learning-based dose prediction for pancreatic stereotactic body radiation therapy using dual pyramid adversarial network.

Treatment planning for pancreatic cancer stereotactic body radiation therapy (SBRT) is very challenging owing to vast spatial variations and close proximity of many organs-at-risk. Recently, deep learning (DL) based methods have been applied in dose prediction tasks of various treatment sites with the aim of relieving planning challenges. However, its effectiveness on pancreatic cancer SBRT is yet to be fully explored due to limited investigations in the literature. This study aims to further current knowledge in DL-based dose prediction tasks by implementing and demonstrating the feasibility of a new dual pyramid networks (DPNs) integrated DL model for predicting dose distributions of pancreatic SBRT. The proposed framework is composed of four parts: CT-only feature pyramid network (FPN), contour-only FPN, late fusion network and an adversarial network. During each phase of the network, combination of mean absolute error, gradient difference error, histogram matching, and adversarial loss is used for supervision. The performance of proposed model was demonstrated for pancreatic cancer SBRT plans with doses prescribed between 33 and 50 Gy across as many as three planning target volumes (PTVs) in five fractions. Five-fold cross validation was performed on 30 patients, and another 20 patients were used as holdout tests of trained model. Predicted plans were compared with clinically approved plans through dose volume parameters and two-paired t-test. For the same sets, our results were compared with three different DL architectures: 3D U-Net, 3D U-Net with adversarial learning, and DPN without adversarial learning. The proposed framework was able to predict 87% and 91% of clinically relevant dose parameters for cross validation sets and holdout sets, respectively, without any significant differences (P > 0.05). Dose distribution predicted by our framework was also able to predict the intentional hotspots as feature characteristics of SBRT plans. Our method achieved higher correlation coefficients with the ground truth in 22/26, 24/26 and 20/26 dose volume parameters compared to the network without adversarial learning, 3D U-Net, and 3D U-Net with adversarial learning, respectively. Overall, the proposed model was able to predict doses to cases with both single and multiple PTVs. In conclusion, the DPN integrated DL model was successfully implemented, and demonstrated good dose prediction accuracy and dosimetric characteristics for pancreatic cancer SBRT.