This study evaluated the feasibility of using artificial intelligence (AI) segmentation software for volume-modulated arc therapy (VMAT) prostate planning in conjunction with knowledge-based planning to facilitate a fully automated workflow. Two commercially available AI software programs, Radformation AutoContour (Radformation, New York, NY) and Siemens AI-Rad Companion (Siemens Healthineers, Malvern, PA) were used to auto-segment the rectum, bladder, femoral heads, and bowel bag on 30 retrospective clinical cases (10 intact prostate, 10 prostate bed, and 10 prostate and lymph node). Physician-segmented target volumes were transferred to AI structure sets. In-house RapidPlan models were used to generate plans using the original, physician-segmented structure sets as well as Radformation and Siemens AI-generated structure sets. Thus, there were three plans for each of the 30 cases, totaling 90 plans. Following RapidPlan optimization, planning target volume (PTV) coverage was set to 95%. Then, the plans optimized using AI structures were recalculated on the physician structure set with fixed monitor units. In this way, physician contours were used as the gold standard for identifying any clinically relevant differences in dose distributions. One-way analysis of variation (ANOVA) was used for statistical analysis. No statistically significant differences were observed across the three sets of plans for intact prostate, prostate bed, or prostate and lymph nodes. The results indicate that an automated volumetric modulated arc therapy (VMAT) prostate planning workflow can consistently achieve high plan quality. However, our results also show that small but consistent differences in contouring preferences may lead to subtle differences in planning results. Therefore, the clinical implementation of auto-contouring should be carefully validated.
With limited high-level evidence, we carried out a comparative effectiveness study for the effect of proton beam therapy (PBT) on overall survival compared to external-beam radiotherapy (EBRT) and brachytherapy (BT) among patients with localized prostate cancer using a national database. PBT was associated with a significant overall survival benefit compared to EBRT and had a similar performance as BT.
Background:
There are few comparative outcomes data regarding the therapeutic delivery of proton beam therapy (PBT) versus the more widely used photon-based external-beam radiation (EBRT) and brachytherapy (BT). We evaluated the impact of PBT on overall survival (OS) compared to EBRT or BT on patients with localized prostate cancer.
Patients and Methods:
The National Cancer Data Base (NCDB) was queried for 2004–2015. Men with clinical stage T1–3, N0, M0 prostate cancer treated with radiation, without surgery or chemotherapy, were included. OS, the primary clinical outcome, was fit by Cox proportional hazard model. Propensity score matching was implemented for covariate balance.
Results:
There were 276,880 eligible patients with a median follow-up of 80.9 months. A total of 4900 (1.8%) received PBT, while 158,111 (57.1%) received EBRT and 113,869 (41.1%) BT. Compared to EBRT and BT, PBT patients were younger and were less likely to be in the high-risk group. On multivariable analysis, compared to PBT, men had worse OS after EBRT (adjusted hazard ratio [HR] = 1.72; 95% confidence interval [CI], 1.51–1.96) or BT (adjusted HR = 1.38; 95% CI, 1.21–1.58). After propensity score matching, the OS benefit of PBT remained significant compared to EBRT (HR = 1.64; 95% CI, 1.32–2.04) but not BT (adjusted HR = 1.18; 95% CI, 0.93–1.48). The improvement in OS with PBT was most prominent in men ≤ 65 years old with low-risk disease compared to other subgroups (interaction P < .001).
Conclusion:
In this national data set, PBT was associated with a significant OS benefit compared to EBRT, and with outcomes similar to BT. These results remain to be validated by ongoing prospective trials.
Postmenopausal women often suffer from vaginal symptoms associated with atrophic vaginitis. Additionally, gynecologic cancer survivors may live for decades with additional, clinically significant, persistent vaginal toxicities caused by cancer therapies, including pain, dyspareunia, and sexual dysfunction. The vaginal microbiome (VM) has been previously linked with vaginal symptoms related to menopause (i.e. dryness). Our previous work showed that gynecologic cancer patients exhibit distinct VM profiles from healthy women, with low abundance of lactobacilli and prevalence of multiple opportunistic pathogenic bacteria. Here we explore the association between the dynamics and structure of the vaginal microbiome with the manifestation and persistence of vaginal symptoms, during one year after completion of cancer therapies, while controlling for clinical and sociodemographic factors. We compared cross-sectionally the vaginal microbiome in 134 women, 64 gynecologic patients treated with radiotherapy and 68 healthy controls, and we longitudinally followed a subset of 52 women quarterly (4 times in a year: pre-radiation therapy, 2, 6 and 12 months post-therapy). Differences among the VM profiles of cancer and healthy women were more pronounced with the progression of time. Cancer patients had higher diversity VMs and a variety of vaginal community types (CTs) that are not dominated by Lactobacilli, with extensive VM variation between individuals. Additionally, cancer patients exhibit highly unstable VMs (based on Bray-Curtis distances) compared to healthy controls. Vaginal symptoms prevalent in cancer patients included vaginal pain (40%), hemorrhage (35%), vaginismus (28%) and inflammation (20%), while symptoms such as dryness (45%), lack of lubrication (33%) and dyspareunia (32%) were equally or more prominent in healthy women at baseline. However, 24% of cancer patients experienced persistent symptoms at all time points, as opposed to 12% of healthy women. Symptom persistence was strongly inversely correlated with VM stability; for example, patients with persistent dryness or abnormally high pH have the most unstable microbiomes. Associations were identified between vaginal symptoms and individual bacterial taxa, including: Prevotella with vaginal dryness, Delftia with pain following vaginal intercourse, and Gemillaceaea with low levels of lubrication during intercourse. Taken together our results indicate that gynecologic cancer therapy is associated with reduced vaginal microbiome stability and vaginal symptom persistence.
Purpose: Treatment with long-term androgen deprivation therapy (ADT) and radiation therapy (RT) is the nonsurgical standard-of-care for patients with high- or very high-risk prostate cancer (HR-PC), but the optimal timing between ADT and RT initiation is unknown. We evaluate the influence of timing between ADT and RT on outcomes in patients with HR-PC using a large national cancer database. Methods and Materials: Data for patients with clinical T1-T4 N0, M0, National Cancer Comprehensive Network HR-PC who were treated with definitive external RT (≥60 Gy) and ADT starting either before or within 14 days after RT start were extracted from the National Cancer Database (2004-2015). Patients were grouped on the basis of ADT initiation: (1) >11 weeks before RT, (2) 8 to 11weeks before RT, and (3) <8 weeks before RT. Kaplan-Meier, propensity score matching, and multivariable Cox proportional hazards were performed to evaluate overall survival (OS). Results: With a median follow-up of 68.9 months, 37,606 patients with HR-PC were eligible for analysis: 13,346 (35.5%) with >11 weeks of neoadjuvant ADT, 11,456 (30.5%) with 8 to 11 weeks of neoadjuvant ADT; and 12,804 (34%) patients with <8 weeks of neoadjuvant ADT. The unadjusted 10-year OS rates for >11 weeks, 8 to 11 weeks, and <8 weeks neoadjuvant ADT groups were 49.9%, 51.2%, and 46.9%, respectively (P =.002). On multivariable and inverse probability of treatment weighting analyses, there was a significant OS advantage for patients in the 8 to 11 weeks neoadjuvant ADT group (adjusted hazard ratio 0.90; 95% confidence interval, 0.86-0.95; P <.001) but not the >11 weeks group. Conclusions: Neoadjuvant ADT initiation 8 to 11 weeks before RT is associated with significantly improved OS compared with shorter neoadjuvant ADT duration. Although prospective validation is warranted, this analysis is the largest retrospective study suggesting an influence of timing between ADT and RT initiation in HR-PC.
Background: Two large randomized trials, CALGB 9343 and PRIME II, support omission of radiotherapy after breast conserving surgery (BCS) in elderly women with favorable-risk early stage breast cancer intending to take endocrine therapy. However, patients with grade 3 histology were underrepresented on these trials. We hypothesized that high-grade disease may be unsuitable for treatment de-escalation and report the oncologic outcomes for elderly women with favorable early stage breast cancer treated with BCS with or without radiotherapy. Materials and Methods: The Surveillance, Epidemiology, and End Results database was queried for women between 70 and 79 years of age with invasive ductal carcinoma diagnosed between 1998 and 2007. This cohort was narrowed to women with T1mic-T1c, N0, estrogen receptor-positive, invasive ductal carcinoma treated with BCS with or without external beam radiation (EBRT). The primary endpoints were 5- and 10-year cause-specific survival (CSS). Univariate and multivariate analyses were performed. Propensity-score matching of T-stage, year of diagnosis, and age was utilized to reduce selection bias while comparing treatment arms within the grade 3 subgroup. Results: A total of 12,036 women met inclusion criteria, and the median follow-up was 9.4 years. EBRT was omitted in 22% of patients, including 21% with grade 3 disease. Patients in the EBRT cohort were slightly younger (median, 74 vs. 75 years; P < .01) and had fewer T1a tumors (11% vs. 13%; P = .02). Histologic grades 1, 2, and 3 comprised 36%, 50%, and 14% of the cohort, respectively, and there were no differences in EBRT utilization by grade. Utilization of EBRT decreased following the publication of the CALGB trial in 2004 decreasing from 82% to 85% in 1998 to 2000 to 73% to 75% in 2005 to 2007 (P < .01). Unadjusted outcomes showed that in grade 1 disease, there were no differences in CSS with or without EBRT at 5 (99%) and 10 years (95%-96%). EBRT was associated with an improvement in CSS in grade 2 histology at 5 years (97% vs. 98%) and 10 years (92% vs. 95%) (P = .004). The benefit was more pronounced in grade 3 disease with CSS increasing from 93% to 96% at 5 years and from 87% to 92% at 10 years (P = .02) with EBRT. In the grade 3 subgroup, propensity-score matching confirmed EBRT was associated with superior CSS compared with surgery alone (hazard ratio, 0.58; 95% confidence interval, 0.34-0.98; P = .043). Conclusion: In this database analysis, omission of radiotherapy after BCS in elderly women with favorable-risk, early stage, grade 3 breast cancer was associated with inferior CSS. Further prospective data in this patient population are needed to confirm our findings and conclusions. Grade 3 disease was not well-represented in clinical trials investigating the omission of radiotherapy following breast conserving surgery in elderly women with early stage breast cancer. This Surveillance, Epidemiology, and End Results analysis of 12,036 women aged 70 to 79 years with T1N0, estrogen receptor-positive, invasive ductal carcinoma found that women with grade 3 disease had both an overall survival and breast cancer-specific mortality benefit with radiotherapy.
Purpose/Objective: Given the rarity of vulvar cancer, data on the incidence of acute and late severe toxicity and patients’ symptom burden from radiotherapy (RT) are lacking. Materials/Methods: This multi-center, single-institution study included patients with vulvar squamous cell carcinoma treated with curative intent RT between 2009 and 2020. Treatment-related acute and late grade ≥ 3 toxicities and late patient subjective symptoms (PSS) were recorded. Results: Forty-two patients with predominantly stage III/IV disease (n = 25, 59.5 %) were treated with either definitive (n = 25, 59.5 %) or adjuvant (n = 17, 40.5 %) external beam RT to a median dose of 64 Gy and 59.4 Gy, respectively. Five patients received a brachytherapy boost with a median total dose of 84.3 Gy in 2 Gy-equivalent dose (EQD2). Intensity-modulated RT was used in 37 (88.1 %) of patients, and 25 patients (59.5 %) received concurrent chemotherapy. Median follow-up was 27 months. Acute grade ≥ 3 toxicity occurred in 17 patients (40.5 %), including 13 (31.0 %) acute grade 3 skin events. No factors, including total RT dose (p = 0.951), were associated with acute skin toxicity. Eleven (27.5 %) patients developed late grade ≥ 3 toxicity events, including 10 (23.8 %) late grade ≥ 3 skin toxicity events. Patients with late grade ≥ 3 skin toxicity had a higher mean body-mass index (33.0 vs 28.2 kg/m2; p = 0.009). Common late PSS included vaginal pain (n = 15, 35.7 %), skin fibrosis (n = 10, 23.8 %), and requirement of long-term opiates (n = 12, 28.6 %). Conclusion: RT for vulvar cancer is associated with considerable rates of severe acute and late toxicity and PSS burden. Larger studies are needed to identify risk factors, explore toxicity mitigation strategies, and assess patient-reported outcomes.
Purpose: High-dose-rate (HDR) brachytherapy is an established technique to be used as monotherapy option or focal boost in conjunction with external beam radiation therapy (EBRT) for treating prostate cancer. Radiation source path reconstruction is a critical procedure in HDR treatment planning. Manually identifying the source path is labor intensive and time inefficient. In recent years, magnetic resonance imaging (MRI) has become a valuable imaging modality for image-guided HDR prostate brachytherapy due to its superb soft-tissue contrast for target delineation and normal tissue contouring. The purpose of this study is to investigate a deep-learning-based method to automatically reconstruct multiple catheters in MRI for prostate cancer HDR brachytherapy treatment planning. Methods: Attention gated U-Net incorporated with total variation (TV) regularization model was developed for multi-catheter segmentation in MRI. The attention gates were used to improve the accuracy of identifying small catheter points, while TV regularization was adopted to encode the natural spatial continuity of catheters into the model. The model was trained using the binary catheter annotation images offered by experienced physicists as ground truth paired with original MRI images. After the network was trained, MR images of a new prostate cancer patient receiving HDR brachytherapy were fed into the model to predict the locations and shapes of all the catheters. Quantitative assessments of our proposed method were based on catheter shaft and tip errors compared to the ground truth. Results: Our method detected 299 catheters from 20 patients receiving HDR prostate brachytherapy with a catheter tip error of 0.37 ± 1.68 mm and a catheter shaft error of 0.93 ± 0.50 mm. For detection of catheter tips, our method resulted in 87% of the catheter tips within an error of less than ± 2.0 mm, and more than 71% of the tips can be localized within an absolute error of no >1.0 mm. For catheter shaft localization, 97% of catheters were detected with an error of <2.0 mm, while 63% were within 1.0 mm. Conclusions: In this study, we proposed a novel multi-catheter detection method to precisely localize the tips and shafts of catheters in three-dimensional MRI images of HDR prostate brachytherapy. It paves the way for elevating the quality and outcome of MRI-guided HDR prostate brachytherapy.
Purpose: Anal cancer affects a disproportionate percentage of persons infected with human immunodeficiency virus (HIV). We analyzed a cohort of patients with HIV and anal cancer who received modern radiation therapy (RT) and concurrent chemotherapy to assess whether certain factors are associated with poor oncologic outcomes. Patients and Methods: We performed a retrospective chart review of 75 consecutive patients with HIV infection and anal cancer who received definitive chemotherapy and RT from 2008 to 2018 at a single academic institution. Local recurrence, overall survival, changes in CD4 counts, and toxicities were investigated. Results: Most patients were male (92%) with large representation from Black patients (77%). The median pretreatment CD4 count was 280 cells/mm3, which was persistently lower at 6 and 12 months’ posttreatment, 87 cells/mm3 and 182 cells/mm3, respectively (P <.001). Most (92%) patients received intensity modulated RT; median dose was 54 Gy (Range, 46.8-59.4 Gy). At a median follow-up 5.4 years (Range, 4.37-6.21 years), 20 (27%) patients had disease recurrence and 10 (13%) had isolated local failures. Nine patients died due to progressive disease. In multivariable analysis, clinically node negative involvement was significantly associated with better overall survival (hazard ratio, 0.39; 95% confidence interval, 0.16-1.00, P =.049). Acute grade 2 and 3 skin toxicities were common, at 83% and 19%, respectively. Acute grade 2 and 3 gastrointestinal toxicities were 9% and 3%, respectively. Acute grade 3 hematologic toxicity was 20%, and one grade 5 toxicity was reported. Several late grade 3 toxicities persisted: gastrointestinal (24%), skin (17%), and hematologic (6%). Two late grade 5 toxicities were noted. Conclusions: Most patients with HIV and anal cancer did not experience local recurrence; however, acute and late toxicities were common. CD4 counts at 6 and 12 months’ posttreatment remained lower than pretreatment CD4 counts. Further attention to treatment of the HIV-infected population is needed.
Accurate tracking of anatomic landmarks is critical for motion management in liver radiation therapy. Ultrasound (US) is a safe, low-cost technology that is broadly available and offer real-time imaging capability. This study proposed a deep learning-based tracking method for the US image-guided radiation therapy. The proposed cascade deep learning model is composed of an attention network, a mask region-based convolutional neural network (mask R-CNN), and a long short-term memory (LSTM) network. The attention network learns a mapping from an US image to a suspected area of landmark motion in order to reduce the search region. The mask R-CNN then produces multiple region-of-interest proposals in the reduced region and identifies the proposed landmark via three network heads: bounding box regression, proposal classification, and landmark segmentation. The LSTM network models the temporal relationship among the successive image frames for bounding box regression and proposal classification. To consolidate the final proposal, a selection method is designed according to the similarities between sequential frames. The proposed method was tested on the liver US tracking datasets used in the medical image computing and computer assisted interventions 2015 challenges, where the landmarks were annotated by three experienced observers to obtain their mean positions. Five-fold cross validation on the 24 given US sequences with ground truths shows that the mean tracking error for all landmarks is 0.65 ± 0.56 mm, and the errors of all landmarks are within 2 mm. We further tested the proposed model on 69 landmarks from the testing dataset that have the similar image pattern with the training pattern, resulting in a mean tracking error of 0.94 ± 0.83 mm. The proposed deep-learning model was implemented on a graphics processing unit (GPU), tracking 47-81 frames s−1. Our experimental results have demonstrated the feasibility and accuracy of our proposed method in tracking liver anatomic landmarks using US images, providing a potential solution for real-time liver tracking for active motion management during radiation therapy.
Objective. CBCTs in image-guided radiotherapy provide crucial anatomy information for patient setup and plan evaluation. Longitudinal CBCT image registration could quantify the inter-fractional anatomic changes, e.g. tumor shrinkage, and daily OAR variation throughout the course of treatment. The purpose of this study is to propose an unsupervised deep learning-based CBCT-CBCT deformable image registration which enables quantitative anatomic variation analysis.Approach.The proposed deformable registration workflow consists of training and inference stages that share the same feed-forward path through a spatial transformation-based network (STN). The STN consists of a global generative adversarial network (GlobalGAN) and a local GAN (LocalGAN) to predict the coarse- and fine-scale motions, respectively. The network was trained by minimizing the image similarity loss and the deformable vector field (DVF) regularization loss without the supervision of ground truth DVFs. During the inference stage, patches of local DVF were predicted by the trained LocalGAN and fused to form a whole-image DVF. The local whole-image DVF was subsequently combined with the GlobalGAN generated DVF to obtain the final DVF. The proposed method was evaluated using 100 fractional CBCTs from 20 abdominal cancer patients in the experiments and 105 fractional CBCTs from a cohort of 21 different abdominal cancer patients in a holdout test.Main Results. Qualitatively, the registration results show good alignment between the deformed CBCT images and the target CBCT image. Quantitatively, the average target registration error calculated on the fiducial markers and manually identified landmarks was 1.91 ± 1.18 mm. The average mean absolute error, normalized cross correlation between the deformed CBCT and target CBCT were 33.42 ± 7.48 HU, 0.94 ± 0.04, respectively.Significance. In summary, an unsupervised deep learning-based CBCT-CBCT registration method is proposed and its feasibility and performance in fractionated image-guided radiotherapy is investigated. This promising registration method could provide fast and accurate longitudinal CBCT alignment to facilitate inter-fractional anatomic changes analysis and prediction.