Patient-Specific 3D CT Images Reconstruction from 2D KV Images Via Vision Transformer-Based Deep-Learning

PURPOSE/OBJECTIVE(S)

In some proton therapy facilities, patient alignment relies on two 2D orthogonal kV images, taken at fixed, oblique angles, as no 3D on-the-bed-imaging is available. The visibility of the tumor in kV images is limited since the patient's 3D anatomy is projected onto a 2D plane, especially when the tumor is behind a high-density structure such as bone. This can lead to a large patient setup error. A solution to this problem is to reconstruct the 3D CT image from the kV images obtained in the treatment position.

MATERIALS/METHODS

An asymmetric autoencoder-like network built with vision-transformer blocks was developed. The data was collected from a head and neck patient: 2 orthogonal kV images (1024X1024 voxels), 1 3D CT with padding (512X512X512) acquired from the in-room CT-on-rails before kVs were taken and 2 digitally-reconstructed-radiograph (DRR) images (512X512) based on the CT. We resampled kV images every 8 voxels and DRR and CT every 4 voxels, thus formed a dataset consisting of 262,144 samples, in which the images had a dimension of 128 for each direction. The value of each voxel in CT was normalized to range 0-1 with a uniform shift of 1000 and a denominator of 4000. For kV and DRR, we ranked all voxels value in an ascending order and normalized the values of the first 80% voxels to range 0-0.8 and the rest to range 0.8-1, thus yielding a quasi-Gaussian distribution, which was favorable by the deep neural networks. We further cropped kV and DRR images with a self-supervised bitmap based on the voxels' gradients. In training, both kV and DRR were utilized, and the encoder was encouraged to learn the same feature maps for kV images and its corresponding DRR images with mean-absolute-error (MAE) as the similarity loss. Then the decoder would reconstruct the 3D CT image from the feature maps of the kV images with the CT-on-rails as ground-truth (gCT) and MAE as the reconstruction loss. In testing, only independent kV images were used. The full-size synthetic CT (sCT) was achieved by concatenating the sCTs generated by the model according to their spatial information. The image quality of the sCT was evaluated using MAE and per-voxel-absolute-CT-number-difference volume histogram (CDVH). The proposed network was implemented with PyTorch deep learning library and both distributed data parallel (DDP) and automatic mixed precision (AMP) were applied to saving memory and accelerating the training speed. We used the AdamW optimizer with β1 = 0.9 and β2 = 0.999 and a cosine annealing learning rate scheduler with an initial learning of 1e-7 and 20 warm-up epochs.

RESULTS

The model achieved a MAE of <40HU and the CDVH showed that <5% of the voxels had a per-voxel-absolute-CT-number-difference larger than 185HU. The profile of a typical gCT slice and its corresponding sCT slice exhibited a high agreement, indicating the high similarity between the gCT and sCT.

CONCLUSION

A patient-specific vision-transformer-based network was developed and shown to be accurate and efficient to reconstruct 3D CT images from kV images.

Physician and Patient-Reported Outcomes of a Phase III Trial of Ultra-Hypofractionated vs. Moderate Hypofractionated Radiotherapy to the Whole Breast after Breast-Conserving Surgery

PURPOSE/OBJECTIVE(S)

To report a final analysis evaluating physician and patient-reported outcomes of early breast cancer patients receiving moderate hypofractionation or ultra-hypofractionated whole breast radiotherapy (RT).

MATERIALS/METHODS

Between April 4, 2018, and February 11, 2020, patients with localized breast cancer (T1-T3, N0-N1, and M0) managed with breast-conserving surgery (BCS) were enrolled. Patients were randomized to receive whole breast RT with moderate hypofractionation to 40 Gy in 15 fractions (Arm A) or ultra-hypofractionation to 25 Gy in 5 fractions (Arm B). An optional concurrent integrated boost to 48 Gy on Arm A or 30 Gy on Arm B was allowed. Early toxicity (<3 months), late toxicity (> 3 months), quality of life (QOL), cosmesis, Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE), and deterioration of cosmesis were analyzed.

RESULTS

One hundred and seven patients were randomized to moderate hypofractionation (n = 54) or ultra-hypofractionation (n = 53). With a median follow-up of 36 months, no significant differences in patient characteristics were noted between the two arms. There were no patients with a grade ≥3 or higher toxicity. Grade 2 toxicities were 7.4% in Arm A and 7.5% in Arm B, and primarily consisted of radiation dermatitis (6 patients), fibrosis (1 patient) and lymphedema (1 patient). The average Harvard Cosmesis score and overall QoL were similar between arms at all time points, with no patients developing cosmetic deterioration. Patient-reported moderate to severe radiation skin burns were more commonly reported in Arm A (21.05%) vs. Arm B (6.25%) at the end of treatment (EOT) (p = 0.078). At EOT, patients receiving moderate hypofractionation had higher mean toxicity scores in breast tenderness (2.66 vs. 1.5, p = 0.018), skin flaking or peeling (0.63 vs. 0.06, p = 0.035), blistering (0.74 vs. 0.06, p = 0.028), pruritis (2.53 vs. 0.87, p < 0.001), erythema (4.24 vs. 2.0, p <0.001), telangiectasias (1.0 vs. 0.28, p = 0.021). Additionally, patients receiving moderate hypofractionation reported significantly worse changes from baseline at EOT in breast tenderness (-2.25 vs. -.86, p = 0.02), telangiectasia (-0.81 vs. 0.18, p = 0.012), skin discoloration (-4.31 vs. -1.04, p < 0.001), skin flaking or peeling (-.55 vs. 0.04, p = 0.053), blistering (-0.82 vs. -0.07, p = 0.033), and pruritus (-2.27 vs. -.67, p = 0.002). There was a return to baseline in all patient-reported breast domains by 3 months (p >0.05) in both arms.

CONCLUSION

Ultra-hypofractionated whole breast irradiation, consisting of 25 Gy in 5 fractions, provided comparable provider assessed toxicity and cosmetic outcomes to 40 Gy in 15 fractions. At the EOT assessment, ultra-hypofractionation had a better patient reported toxicity profile. Our findings provide further evidence to support daily ultra-hypofractionated whole breast radiotherapy as an appropriate treatment option for early-stage breast cancer.

Oncologic Outcomes of Conventionally Fractionated, Hypofractionated, and Stereotactic Body Spot-Scanned Proton Radiation Therapy for Prostate Cancer: The Mayo Clinic Experience

PURPOSE/OBJECTIVE(S)

Spot/pencil beam scanned proton therapy is a relatively new technology with fundamental differences from double scattered or IMRT. We aimed to report the long-term oncologic outcomes of a contemporary prospective series of patients treated with spot-scanned proton therapy (SSPT).

MATERIALS/METHODS

An IRB-approved prospective registry identified patients with prostate cancer treated with proton therapy between January 2016 and December 2018. Descriptive statistics were calculated for all patients. Clinical, demographic, and treatment characteristics were gathered and analyzed. Kaplan-Meier curves were generated to estimate survival and recurrence rates. Outcomes assessed included 5-year overall survival (OS), 5-year local control (LC), biochemical failure (BF), regional and distant failures, and physician-reported adverse events (AEs). Biochemical failure was defined as rise in PSA ≥ 2.0 ng/mL above nadir PSA. Acute and chronic gastrointestinal (GI) and genitourinary (GU) grade 2+ and grade 3+ baseline-adjusted AEs were assigned using CTCAE v5.0. All failures were re-staged with PET C-11 or PSMA.

RESULTS

With a median follow up of 4.4 years (IQR 3.7 - 5), two hundred and eighty-six prostate cancer patients with a median age of 72 (IQR 67.5 - 77) were treated with spot-scanned proton radiation. The median Gleason grade group was 3 (IQR 2 - 4). The median pre-RT PSA was 6.9 ng/mL (IQR 4.3 - 10.5). Median T-stage was T1c. Nearly 64% of all patients were on androgen deprivation therapy at the time of initiating radiation treatment. The median total radiation dose was 79.2 Gy delivered over 44 fractions, 70 Gy over 28 fractions, and 38 Gy over 5 fractions for CF, HF, and SBRT regimens, respectively. The BF rate for all patients was 8.4%. The 5-year LC rates for CF, HF, and SBRT were 100% (95% CI: 100 - 100), 100% (95% CI: 100 - 100), and 97.3% (95% CI: 92.2 - 100), respectively (p = 0.07). Regional recurrences occurred in 12 (4.2%) patients: 8 (5.6%) treated with CF, 2 (2.1%) with HF, and 2 (4.3%) with SBRT (p = 0.62). Distant metastatic failures occurred in 12 patients (4.2%): 5 (3.5%) treated with CF, 7 (7.4%) with HF, and none with SBRT (0%) (p = 0.052). The 5-year OS for patients treated with CF, HF, and SBRT SSPT were 88.2% (95% CI: 81.8 - 95), 86.2% (95% CI: 77.6 - 95.6), and 97.2% (95% CI: 92 - 100), respectively (p = 0.1). Acute and chronic grade 2+ GI baseline-adjusted AEs occurred in 8 (2.8%) and 51 (17.8%) patients, respectively. Acute and chronic grade 3+ GI baseline-adjusted AEs occurred in 3 (1%) and 4 (1.4%) patients, respectively. Acute and chronic grade 2+ GU-related AEs were observed in 72 (25.2%) and 63 (22%) patients, respectively. Acute and chronic grade 3+ GU toxicity was observed in 3 (1%) and 6 (2.1%) patients, respectively.

CONCLUSION

Spot-scanned proton radiation therapy provides high local control rates and excellent oncologic outcomes across different fractionation schedules with low long-term AE rates.

Final Analysis of a Phase III Controlled Randomized Study of Stereotactic Body Proton Therapy or Conventionally Fractionated Proton Therapy for Early Prostate Cancer: PCG GU002

PURPOSE/OBJECTIVE(S)

To determine if stereotactic body proton therapy (SBPT) is non-inferior to conventionally fractionated proton therapy (CFPT) in patients with early prostate cancer.

MATERIALS/METHODS

Multicenter, randomized, controlled, open-label, non-inferiority phase 3 trial that included patients with histologically confirmed low-risk prostate adenocarcinoma defined by Gleason score ≤6, PSA <10 ng/mL, and clinical stage T1-2a N0 M0 by AJCC 7th Ed. Eligible participants were randomly assigned (initially 1:1 and later 2:1 ratio) to CFPT (79.2 Gy in 44 fractions for 9 weeks) or SBPT (38 Gy in 5 fractions for 1 week). Concurrent or adjuvant androgen deprivation therapy was not allowed. The primary endpoint was freedom from failure (FFF) at 2 years, defined as the first occurrence of local, regional, or distant recurrence, biochemical failure by the Phoenix definition (increase of PSA ≥2 ng/mL over the nadir PSA), or the start of salvage therapy including ADT. Secondary endpoints included GI and GU grade ≥2 toxicity according to CTCAE v4 criteria, as well as health-related quality of life (HRQoL) metrics assessed by AUASI and EPIC scores. Non-inferiority would be declared if the 1-sided 95% confidence interval limit for the difference in 2-year FFF rate was below 4.2% between both groups by Clopper-Pearson exact method.

RESULTS

Between November 2010 and September 2020, 133 patients were enrolled and randomly assigned to CFPT (n = 45) or SBPT (n = 88). Median follow-up was 5 years (IQ 3.9-5.2), with the last patient enrolled followed for at least 2 years. The 2-year FFF was 100% for both groups, fulfilling the pre-specified criteria for non-inferiority of SBPT compared to CFPT. By KM estimates, 5-year FFF was 97.4% and 100% (P = 0.1), and the 5-year OS was 97.1% and 95.5% (P = 0.46) for patients treated with CFPT and SBPT, respectively. The cumulative incidence of any grade ≥3 toxicities at 5 years was 0% and 5.7% (P = 0.14) for patients treated with CFPT and SBPT, respectively. The frequency of GI grade ≥2 toxicity at 6 months was of 0% and 2.3% (P = 0.55), and at 2 years was of 6.7% and 3.4% (P = 0.69) for patients treated with CFPT and SBPT, respectively. The frequency of GU grade ≥2 toxicity at 6 months was of 2.2% and 5.7% (P = 0.42), and at 2 years was of 8.9% and 5.7% (P = 0.54) for patients treated with CFPT and SBPT, respectively. Changes in HRQoL scores at 2 years were similar between groups (Table).

CONCLUSION

SBPT is non-inferior to CFPT regarding FFF and associated with similar long-term toxicity rates and HRQoL metric scores.

Sensitivity of the PROMIS-10 for Capturing Radiation-Related Quality of Life Changes

PURPOSE/OBJECTIVE(S)

Patient reported outcomes (PROs) are becoming more common when assessing the effects of radiotherapy (RT). The aim of this study was to assess the sensitivity of the Mental and Physical domains of the Patient-Reported Outcomes Measurement Information System 10 (PROMIS-10) to radiotherapy and determine what predictors were associated with change in quality of life.

MATERIALS/METHODS

Patients, regardless of cancer type, were enrolled on a multi-site prospective registry. Inclusion criteria included curative radiotherapy and completion of the PROMIS-10 prior to treatment (Baseline) and at End of Treatment (EOT). To assess the strongest predictors of change in the T score of mental and physical health, we included 14 demographic characteristics and treatment variables in a multivariable stepwise regression.

RESULTS

A total of 7,586 patients were eligible for the analysis. The median age was 65 (range 18-94), 54% were males, and 94% were white. A majority received photons (62.5%) and the others received protons (37.5%) with an average dose of 52.3 Gy (range 20-80 Gy) over an average of 22.6 fractions (range 1-66). Patient disease sites were sub-grouped into 12 categories: Breast (25.5%), GU (23.0%), H&N (11.1%), CNS (8.5%), Pancreas-Biliary (6.7%), Thoracic (5.7%), Soft Tissue/Bone (5.0%), Esophagus-Gastric (4.7%), Colorectal-Anus (4.4%), Heme/Lymph (2.6%), GYN (1.8%), and Skin/Melanoma (1.0%). For both outcomes, the model selected disease group as an important predictor and it explained the most variance in the outcome compared to the rest of the predictors. When probing the effect of disease group, H&N, Esophagus-Gastric, Skin/Melanoma, and Colorectal-Anus had the largest mean decrease in quality of life for both domains. For mental health, the model also selected radiation type. Patients treated with protons indicated a bigger decrease in mental health compared to patients treated with photons (b = 0.43, 95% CI: -0.01, 0.69). For physical health, the model selected total fractions, ethnicity, and T stage. As number of fractions increased, the physical health change scores became more negative, on average (b = -0.03, 95% CI: -0.05, -0.01). Hispanic/Latino patients indicated a smaller decrease in physical health compared to White (b = -1.50, 95% CI: -2.60, -0.40) and Unknown ethnicity patients (b = -1.82, 95% CI: -3.36, -0.27). Finally, patients with a T stage of 3 or greater indicated a smaller decrease in physical health than patients with a T stage less than 3 (b = 0.76, 95% CI: 0.35, 1.16).

CONCLUSION

The PROMIS-10 did not capture significant change for patients undergoing curative radiotherapy except for patients with Head & Neck, Esophagus-Gastric, Skin, and Colorectal-Anus cancer. Further analyses should explore which patients experience the greatest change in quality of life within disease group.

Treatment of Thymoma and Thymic Carcinoma with Proton Beam Therapy: Outcomes from the Proton Collaborative Group Prospective Registry

PURPOSE/OBJECTIVE(S)

Given the generally long natural history of thymic malignancies, proton beam therapy (PBT) is advocated to minimize the risk of long-term toxicities to mediastinal organs. Adverse events (AE) and long-term clinical outcomes for this population have not been well-characterized.

MATERIALS/METHODS

The Proton Collaborative Group registry (NCT01255748), a multi-institutional prospective database of academic and community proton centers in the US, was queried for patients with thymomas and thymic carcinomas treated with PBT. Patients with recurrent/metastatic disease, non-thymic histology, received either prior or palliative radiotherapy (dose < 40 Gy RBE) were excluded. Overall survival (OS) and local control (LC) were estimated using Kaplan-Meier methods.

RESULTS

A total of 97 patients were identified in the PCG registry. After applying relevant exclusion criteria, 70 patients from 12 proton centers treated from 2011-2021 were included for analysis. Median follow-up length was 16 months. Median age was 58.5 years (IQR 46-63), and 60% were female. 81.4% had a diagnosis of thymoma, and 18.6% thymic carcinoma. 59 patients underwent surgical resection. 11 were treated with definitive PBT, of which 5 received concurrent chemotherapy. Median dose was 54 Gy RBE (range 41.4 - 70 Gy RBE), median number of fractions was 30 (range 21 - 38). 73.4% received pencil beam scanning and 23% uniform scanning PBT. Treatment was overall well-tolerated: a single patient developed grade 4 pneumonitis. Grade 3 AEs were seen in 3 patients - dyspnea, anorexia, and heart failure. Highest grade toxicity experienced was grade 2 for 47.1% and grade 1 for 42.9% of patients. 3-year overall survival (OS) was 82.6% for the entire cohort. 3-year OS was 94% for resected/adjuvant cohort and 35.6% in the non-surgical/definitive cohort. 3-year local control (LC) was 91.7% for the entire cohort. By surgery/margin status, 3-year LC was 96.8% in patients with close or negative margins (a single failure in a patient with close margins), whereas 3-year LC was 55.1% for patients with positive margins/unresectable disease.

CONCLUSION

Thymic malignancies treated with PBT appear to have favorable outcomes, especially in the adjuvant setting, in this cohort representing the largest series of such patients.

Patient-reported outcomes version of the common terminology criteria for adverse events and quality-of-life linear analogue self-assessment in breast cancer patients receiving radiation therapy: single-institution prospective registry

Purpose/objectives

We sought to investigate the impact of patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE) on overall quality-of-life (QOL) employing linear analogue self-assessment (LASA) in breast cancer (BC) patients undergoing radiation therapy (RT).

Materials/methods

All patients treated with RT for BC with curative intent from 2015 to 2019 at our institution were included. Breast specific PRO-CTCAE and overall QOL LASA questionnaires were administered at baseline, end-of-treatment, 3, 6, 12 months, and then annually. Minimal clinically important difference in overall QOL was a 10-point change in LASA. Hypofractionation was any treatment > 2 Gy per fraction. Mixed models for repeated measures were used to determine the association of PRO-CTCAE and overall QOL LASA.

Results

Three hundred thirty-one (331) patients with a median follow-up of 3.1 years (range 0.4–4.9) were included. Average overall QOL LASA scores were 78.5 at baseline, 79.8 at end-of-treatment, 79.8 at 3 months, 77.1 at 6 months, 79.4 at 12 months, and 79.7 at 24 months. On univariate analysis, patients reporting a grade ≥ 3 PRO-CTCAE had, on average, a 10.4-point reduction in overall LASA QOL (p < 0.0001). On multivariate analysis, not being treated with hypofractionation and higher BMI were predictive for worse overall LASA QOL with a 10-point reduction in LASA for patients reporting a grade ≥ 3 PRO-CTCAE (p < 0.0001).

Conclusions

Patients reporting a grade ≥ 3 PRO-CTCAE experienced statistically significant and clinically meaningful deterioration in overall QOL LASA. Hypofractionation improved QOL while higher BMI predicted for worse QOL. PRO-CTCAE should be integrated into future clinical trials.

Abstract OT2-03-03: A randomized trial of 15 fraction vs 25 fraction pencil beam scanning proton radiotherapy after mastectomy in patients requiring regional nodal irradiation

Background: Post-mastectomy radiotherapy improves survival in women with node-positive breast cancer. Pencil beam scanning proton therapy is attractive due to potential to reduce the dose to the heart and lungs compared with traditional photon techniques while improving conformality and limiting skin dose compared with passively scanned proton therapy. The optimal dose and fractionation for pencil-beam scanning proton therapy remains unknown.

Trial Design: This is a multi-center open label phase II randomized controlled trial to determine the safety of 15 fraction vs 25 fraction pencil beam scanning proton radiotherapy after mastectomy in patients requiring regional nodal irradiation.

Eligibility Criteria: Patients ≥ 18 years with primary, non-inflammatory invasive breast cancer who have undergone mastectomy with or without immediate reconstruction and chest wall and regional nodal irradiation planned.

Aims: To determine whether the 24 month complication rate (defined as grade 3 or greater late adverse events, and unplanned surgical intervention in patients who undergo mastectomy with reconstruction) of 15 fraction chest wall and regional node pencil beam scanning proton radiotherapy is acceptable relative to 25 fraction chest wall and regional nodal pencil beam scanning proton radiotherapy and worthy of further investigation.

Statistical methods: The study is designed as a non-inferiority/superiority “hybrid” design using the approach of Freidlin et al. It is Using a one-sided type I error rate of 0.05 (corresponding equivalently to constructing a 1-sided 95% confidence limit), 72 evaluable patients will have 80% power to reject the null hypothesis that the 24-month complication rate in the experimental arm is higher than that of the control arm by more than 10% (i.e. rule out inferiority) under the alternative hypothesis that the complication rate in the experimental arm is 5% less than that of the control arm (i.e. superiority). However, the design will have only 41% power when the two treatment arms are equivalent (i.e. the complication rate is 10% for both arms).

Accrual: The study opened in June 2016. Five of a planned eighty-two patients have been accrued to date.

Citation Format: Mutter RW, Park SS, Hieken TJ, Vargas CE, Mei-Yin PC, Kathryn RJ, Hector VR, Kimberly CS, Elizabeth YS, Daniel VW. A randomized trial of 15 fraction vs 25 fraction pencil beam scanning proton radiotherapy after mastectomy in patients requiring regional nodal irradiation [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr OT2-03-03.