Mean esophageal radiation dose is predictive of the grade of acute esophagitis in lung cancer patients treated with concurrent radiotherapy and chemotherapy

Combined deep learning and radiomics in pretreatment radiation esophagitis prediction for patients with esophageal cancer underwent volumetric modulated arc therapy

PURPOSE

To develop a combined radiomics and deep learning (DL) model in predicting radiation esophagitis (RE) of a grade ≥ 2 for patients with esophageal cancer (EC) underwent volumetric modulated arc therapy (VMAT) based on computed tomography (CT) and radiation dose (RD) distribution images.

MATERIALS AND METHODS

A total of 273 EC patients underwent VMAT were retrospectively reviewed and enrolled from two centers and divided into training (n = 152), internal validation (n = 66), and external validation (n = 55) cohorts, respectively. Radiomic and dosiomic features along with DL features using convolutional neural networks were extracted and screened from CT and RD images to predict RE. The performance of these models was evaluated and compared using the area under curve (AUC) of the receiver operating characteristic curves (ROC).

RESULTS

There were 5 and 10 radiomic and dosiomic features were screened, respectively. XGBoost achieved a best AUC of 0.703, 0.694 and 0.801, 0.729 with radiomic and dosiomic features in the internal and external validation cohorts, respectively. ResNet34 achieved a best prediction AUC of 0.642, 0.657 and 0.762, 0.737 for radiomics based DL model (DLR) and RD based DL model (DLD) in the internal and external validation cohorts, respectively. Combined model of DLD + Dosiomics + clinical factors achieved a best AUC of 0.913, 0.821 and 0.805 in the training, internal, and external validation cohorts, respectively.

CONCLUSION

Although the dose was not responsible for the prediction accuracy, the combination of various feature extraction methods was a factor in improving the RE prediction accuracy. Combining DLD with dosiomic features was promising in the pretreatment prediction of RE for EC patients underwent VMAT.

Enhanced prediction of postoperative radiotherapy-induced esophagitis in non-small cell lung cancer: Dosiomic model development in a real-world cohort and validation in the PORT-C randomized controlled trial

BACKGROUND

Radiotherapy-induced esophagitis (RE) diminishes the quality of life and interrupts treatment in patients with non-small cell lung cancer (NSCLC) undergoing postoperative radiotherapy. Dosimetric models showed limited capability in predicting RE. We aimed to develop dosiomic models to predict RE.

METHODS

Models were trained with a real-world cohort and validated with PORT-C randomized controlled trial cohort. Patients with NSCLC undergoing resection followed by postoperative radiotherapy between 2004 and 2015 were enrolled. The endpoint was grade ≥2 RE. Esophageal three-dimensional dose distribution features were extracted using handcrafted and convolutional neural network (CNN) methods, screened using an entropy-based method, and selected using minimum redundancy and maximum relevance. Prediction models were built using logistic regression. The areas under the receiver operating characteristic curve (AUC) and precision-recall curve were used to evaluate prediction model performance. A dosimetric model was built for comparison.

RESULTS

A total of 190 and 103 patients were enrolled in the training and validation sets, respectively. Using handcrafted and CNN methods, 107 and 4096 features were derived, respectively. Three handcrafted, four CNN-extracted and three dosimetric features were selected. AUCs of training and validation sets were 0.737 and 0.655 for the dosimetric features, 0.730 and 0.724 for handcrafted features, and 0.812 and 0.785 for CNN-extracted features, respectively. Precision-recall curves revealed that CNN-extracted features outperformed dosimetric and handcrafted features.

CONCLUSIONS

Prediction models may identify patients at high risk of developing RE. Dosiomic models outperformed the dosimetric-feature model in predicting RE. CNN-extracted features were more predictive but less interpretable than handcrafted features.

Analysis of deep inspiration breath-hold technique to improve dosimetric and clinical advantages in postoperative intensity-modulated radiation therapy for thymomas

Background

Radiation therapy is one of the essential treatment modalities for invasive thymomas. Clinically, respiratory motion poses a challenge for the radiotherapy of thoracic tumors. One method to address this issue is to train patients to hold their breath at the end of deep inspiration. The purpose of this retrospective cohort study was to investigate the dosimetric and clinical advantages of the deep inspiration breath-hold (DIBH) technique in postoperative intensity-modulated radiation therapy (IMRT) for thymomas.

Methods

Thymoma patients undergoing postoperative IMRT were included. Each patient underwent two computed tomography (CT) scans, one under free breath (FB) and the other under DIBH. Dosimetric parameters of organs at risk (OARs) were evaluated in three series plans. Dose analysis and volume comparisons were conducted during FB-3 mm (FB with 3 mm internal target volume margin), FB-10 mm (FB with 10 mm internal target volume margin), and DIBH and compared using a paired sample Student’s t-test. Normal tissue complication probabilities (NTCP) for lungs and heart were calculated and compared.

Results

The total lung volume significantly increased by 31% (4,216±198 vs. 2,884±166 mL) and the heart volume reduced by 12% (552±25 vs. 636±35 mL) between DIBH acquisitions compared to FB. A significant improvement was observed in all the dosimetric parameters (D_mean, V20, V5) of the lung on DIBH compared to FB-3 mm (54%±2.85% vs. 47%±2.90%, P<0.001; 15%±1.37% vs. 12%±1.32%, P=0.004; and 10.28±0.58 vs. 8.76±0.57 Gy, P<0.001, respectively), as well as in the D_max and D2% of the esophagus and spine. The lung volume increment was related to a reduction in the mean dose of lungs, with a correlation coefficient of r=0.27, P=0.03. The NTCP values for pneumonitis significantly reduced with DIBH compared to the FB state (0.6% vs. 1.1%, P<0.001).

Conclusions

The radiation dose to the OARs can be significantly reduced by using the DIBH technique in postoperative IMRT for thymomas. The increased volume of lungs using DIBH acquisitions can significantly reduce the incidence of pneumonitis.

Predictive Modeling of Thoracic Radiotherapy Toxicity and the Potential Role of Serum Alpha-2-Macroglobulin

Background: To investigate the impact of alpha-2-macroglobulin (A2M), a suspected intrinsic radioprotectant, on radiation pneumonitis and esophagitis using multifactorial predictive models. Materials and Methods: Baseline A2M levels were obtained for 258 patients prior to thoracic radiotherapy (RT). Dose-volume characteristics were extracted from treatment plans. Spearman's correlation (Rs) test was used to correlate clinical and dosimetric variables with toxicities. Toxicity prediction models were built using least absolute shrinkage and selection operator (LASSO) logistic regression on 1,000 bootstrapped datasets. Results: Grade ≥2 esophagitis and pneumonitis developed in 61 (23.6%) and 36 (14.0%) patients, respectively. The median A2M level was 191 mg/dL (range: 94-511). Never/former/current smoker status was 47 (18.2%)/179 (69.4%)/32 (12.4%). We found a significant negative univariate correlation between baseline A2M levels and esophagitis (Rs = -0.18/p = 0.003) and between A2M and smoking status (Rs = 0.13/p = 0.04). Further significant parameters for grade ≥2 esophagitis included age (Rs = -0.32/p < 0.0001), chemotherapy use (Rs = 0.56/p < 0.0001), dose per fraction (Rs = -0.57/p < 0.0001), total dose (Rs = 0.35/p < 0.0001), and several other dosimetric variables with Rs > 0.5 (p < 0.0001). The only significant non-dosimetric parameter for grade ≥2 pneumonitis was sex (Rs = -0.32/p = 0.037) with higher risk for women. For pneumonitis D15 (lung) (Rs = 0.19/p = 0.006) and D45 (heart) (Rs = 0.16/p = 0.016) had the highest correlation. LASSO models applied on the validation data were statistically significant and resulted in areas under the receiver operating characteristic curve of 0.84 (esophagitis) and 0.78 (pneumonitis). Multivariate predictive models did not require A2M to reach maximum predictive power. Conclusion: This is the first study showing a likely association of higher baseline A2M values with lower risk of radiation esophagitis and with smoking status. However, the baseline A2M level was not a significant risk factor for radiation pneumonitis.

Acute radiation oesophagitis associated with 2-deoxy-2-[18F]fluoro-d-glucose uptake on positron emission tomography/CT during chemo-radiation therapy in patients with non-small-cell lung cancer

INTRODUCTION

Acute radiation oesophagitis (ARO) is frequently experienced by patients receiving concurrent chemo-radiation therapy (cCRT) for non-small-cell lung cancer (NSCLC). We investigated ARO symptoms (CTCAE v3.0), radiation dose and oesophageal FDG PET/CT uptake.

METHOD

Candidates received cCRT (60 Gy, 2 Gy/fx) and sequential FDG PET/CT (baseline FDG₀ , FDG_wk2 and FDG_wk4 ). Mean and maximum standardized uptake value (SUVmean and SUVmax) and radiation dose (O_mean and O_max ) were calculated within the whole oesophagus and seven sub-regions (5-60 Gy).

RESULTS

Forty-four patients underwent FDG₀ and FDG_wk2 , and 41 (93%) received FDG_wk4 , resulting in 129 PET/CT scans for analysis. Of 29 (66%) patients with ≥ grade 2 ARO, SUVmax (mean ± SD) increased from FDG₀ to FDG_wk4 (3.06 ± 0.69 to 3.83 ± 1.27, P = 0.0019) and FDG_wk2 to FDG_wk4 (3.10 ± 0.75 to 3.83 ± 1.27, P = 0.0046). Radiation dose (mean ± SD) was higher in grade ≥2 patients; O_mean (47.5 ± 20 vs 53.9 ± 10.2, P = 0.0061), O_max (13.7 ± 9.6 vs 20.1 ± 10.6, P = 0.0009) and V40 Gy (8.0 ± 8.2 vs 11.9 ± 7.3, P = 0.0185).

CONCLUSIONS

FDG_wk4 SUVmax and radiation dose were associated with ≥ grade 2 ARO. Compared to subjective assessments, future interim FDG PET/CT acquired for disease response assessment may also be utilized to objectively characterize ARO severity and image-guided oesophageal dose constraints.

Predictors for severe acute esophagitis in lung cancer patients treated with chemoradiotherapy: a systematic review

PURPOSE

To identify the risk factors for severe acute esophagitis (AE) in lung cancer patients undergoing chemoradiotherapy (CRT).

METHODS

Articles from PubMed, EMBASE, and the Cochrane Library were searched in August 2015. Articles reporting studies of the predictors for severe AE in lung cancer patients after CRT were included. Study quality was assessed using a modified quality assessment tool that was designed previously for an observational study. The effects of studies were combined with the study quality score using a best-evidence synthesis model. Severe AE incidence was also performed using the Metafor package of R-2.11.1.

RESULTS

A total of nine observational studies involving 1641 patients were included. The estimated incidence of severe AE was 14%. According to the best-evidence synthesis criteria, there were two strong-evidence risk factors for severe AE, which were the use of concurrent chemotherapy (CCT) and dose volume histogram (DVH). We also identified four limited-evidence risk factors.

CONCLUSIONS

More attention should be paid to the levels of patients' esophagus function. Although there is no conclusive evidence for severe AE in lung cancer patients after CRT, the above-mentioned factors provide evidence to guide clinicians as to which patients will have severe AE and to choose an optimal prophylactic strategy.

Oesophagus side effects related to the treatment of oesophageal cancer or radiotherapy of other thoracic malignancies

The oesophagus as a serial organ located in the central chest is frequent subject to "incidental" dose application in radiotherapy for several thoracic malignancies including oesophageal cancer itself. Especially due to the radiosensitive mucosa severe radiotherapy induced sequelae can occur, acute oesophagitis and strictures as late toxicity being the most frequent side-effects. In this review we focus on oesophageal side effects derived from treatment of gastrointestinal cancer and secondly provide an overview on oesophageal toxicity from conventional and stereotactic fractionated radiotherapy to the thoracic area in general. Available data on pathogenesis, frequency, onset, and severity of oesophageal side effects are summarized. Whereas for conventional radiotherapy the associations of applied doses to certain volumes of the oesophagus are well described, the tolerance dose to the mediastinal structures for hypofractionated therapy is unknown. The review provides available attempts to predict the risk of oesophageal side effects from dosimetric parameters of SBRT.

Dosimetric evaluation of a simple planning method for improving intensity-modulated radiotherapy for stage III lung cancer

This study aimed to evaluate the dosimetric outcomes of a base-dose-plan-compensation (BDPC) planning method for improving intensity-modulated radiotherapy (IMRT) for stage III lung cancer. For each of the thirteen included patients, three types of planning methods were applied to obtain clinically acceptable plans: (1) the conventional optimization method (CO); (2) a split-target optimization method (STO), in which the optimization objectives were set higher dose for the target with lung density; (3) the BDPC method, which compensated for the optimization-convergence error by further optimization based on the CO plan. The CO, STO and BDPC methods were then compared regarding conformity index (CI), homogeneity index (HI) of the target, organs at risk (OARs) sparing and monitor units (MUs). The BDPC method provided better HI/CI by 54%/7% on average compared to the CO method and by 38%/3% compared to the STO method. The BDPC method also spared most of the OARs by up to 9%. The average MUs of the CO, STO and BDPC plans were 890, 937 and 1023, respectively. Our results indicated that the BDPC method can effectively improve the dose distribution in IMRT for stage III lung cancer, at the expense of more MUs.

Advances in dosimetry and biological predictors of radiation-induced esophagitis

OBJECTIVE

To summarize the research progress about the dosimetry and biological predictors of radiation-induced esophagitis.

METHODS

We performed a systematic literature review addressing radiation esophagitis in the treatment of lung cancer published between January 2009 and May 2015 in the PubMed full-text database index systems.

RESULTS

Twenty-eight eligible documents were included in the final analysis. Many clinical factors were related to the risk of radiation esophagitis, such as elder patients, concurrent chemoradiotherapy, and the intense radiotherapy regimen (hyperfractionated radiotherapy or stereotactic body radiotherapy). The parameters including Dmax, Dmean, V20, V30, V50, and V55 may be valuable in predicting the occurrence of radiation esophagitis in patients receiving concurrent chemoradiotherapy. Genetic variants in inflammation-related genes are also associated with radiation-induced toxicity.

CONCLUSION

Dosimetry and biological factors of radiation-induced esophagitis provide clinical information to decrease its occurrence and grade during radiotherapy. More prospective studies are warranted to confirm their prediction efficacy.

Predicting Radiation Esophagitis Using 18F-FDG PET During Chemoradiotherapy for Locally Advanced Non-Small Cell Lung Cancer

INTRODUCTION

Treatment of locally advanced non-small cell lung cancer with chemoradiotherapy (CRT) is limited by development of toxicity in normal tissue, including radiation esophagitis (RE). Increasingly, (18)F-fluorodeoxyglucose (FDG) positron emission tomography (PET) is being used for adaptive planning. Our aim was to assess changes in esophageal FDG uptake during CRT and relate the changes to the onset and severity of RE.

METHODS

This prospective study in patients with stage II-III non-small cell lung cancer involved serial four-dimensional computed tomography and PET scans during CRT (60-74Gy). RE was recorded weekly using the Common Terminology Criteria for Adverse Events (v4.0), and imaging was performed at weeks 0, 2, 4, and 7. Changes in the esophagus's peak standard uptake value (SUVpeak) were analyzed for each time point and correlated with grade of RE using the Wilcoxon rank-sum test. The volume of esophagus receiving 50 Gy (V50) and volume of esophagus receiving 60 Gy (V60) were correlated with the development of RE, and the C-statistic (area under the curve [AUC]) was calculated to measure predictivity of grade 3 RE.

RESULTS

RE developed in 20 of 27 patients (74%), with grade 3 reached in 6 (22%). A significant percentage increase in SUVpeak in the patients with RE was noted at week 4 (p = 0.01) and week 7 (p = 0.03). For grade 3 RE, a significant percentage increase in SUVpeak was noted at week 2 (p = 0.01) and week 7 (p = 0.03) compared with that for less than grade 3 RE. Median V50 (46.3%) and V60 (33.4%) were significantly higher in patients with RE (p = 0.04). The AUC measurements suggested that the percentage change in SUVpeak at week 2 (AUC = 0.69) and V50 (AUC = 0.67) and V60 (AUC = 0.66) were similarly predictive of grade 3 RE.

CONCLUSIONS

Serial FDG-PET images during CRT show significant increases in SUVpeak for patients in whom RE develops. The changes at week 2 may predict those at risk for the development of grade 3 RE and may be informative for adaptive planning and early intervention.

Timing and intensity of changes in FDG uptake with symptomatic esophagitis during radiotherapy or chemo-radiotherapy

Purpose

To study whether esophageal FDG activity changes by time of mid-course of fractionated radiotherapy (RT), and whether these changes are associated with radiation esophagitis in patients with non-small cell lung cancer (NSCLC).

Methods

Fifty patients with stage I-III NSCLC were enrolled prospectively and, all received ≥60 Gy RT. FDG-PET/CT scans were acquired prior to, and during-RT after delivery of 45 Gy. Normalized standardized uptake values (NSUV), defined by the esophageal maximum SUV relative to intravascular background level in the aortic arch, were sampled in the esophagus at the level of the primary tumor, sternal notch, aortic arch, carina, and gastro-esophageal junction. Symptomatic radiation esophagitis was defined as an event.

Results

Compared to baseline, esophageal NSUV increased significantly during-RT at the level of the primary tumor (1.09 ± 0.05 vs.1.28 ± 0.06, p = 0.001), but did not change at other levels in the esophagus. 16 patients had radiation esophagitis events and these patients had significantly higher during-RT to baseline NSUV ratios than those without esophagitis (1.46 ± 0.12, 95% CI 1.20-1.71; vs. 1.11 ± 0.05, 95% CI 1.01-1.21, p = 0.002). Maximum esophageal dose (p = 0.029), concurrent chemotherapy (p = 0.022) and esophageal FDG PET NSUV ratio (during-RT to baseline, p = 0.007), were independent factors associated with esophagitis and area under curves (AUC) were 0.76, 0.70 and 0.78, respectively. Combining esophageal maximum dose and FDG PET NSUV Ratio at the tumor level increased AUC to 0.85 (p = 0.016).

Conclusion

FDG uptake increased in esophagus during-RT and this increase may predict radiation esphagitis during later course of treatment.