Imaging science and development in modern high-precision radiotherapy

Assessing the generalisability of radiomics features previously identified as predictive of radiation-induced sticky saliva and xerostomia

Background and purpose

While core to the scientific approach, reproducibility of experimental results is challenging in radiomics studies. A recent publication identified radiomics features that are predictive of late irradiation-induced toxicity in head and neck cancer (HNC) patients. In this study, we assessed the generalisability of these findings.

Materials and Methods

The procedure described in the publication in question was applied to a cohort of 109 HNC patients treated with 50-70 Gy in 20-35 fractions using helical radiotherapy although there were inherent differences between the two patient populations and methodologies. On each slice of the planning CT with delineated parotid and submandibular glands, the imaging features that were previously identified as predictive of moderate-to-severe xerostomia and sticky saliva 12 months post radiotherapy (Xer12m and SS12m) were calculated. Specifically, Short Run Emphasis (SRE) and maximum CT intensity (maxHU) were evaluated for improvement in prediction of Xer12m and SS12m respectively, compared to models solely using baseline toxicity and mean dose to the salivary glands.

Results

None of the associations previously identified as statistically significant and involving radiomics features in univariate or multivariate models could be reproduced on our cohort.

Conclusion

The discrepancies observed between the results of the two studies delineate limits to the generalisability of the previously reported findings. This may be explained by the differences in the approaches, in particular the imaging characteristics and subsequent methodological implementation. This highlights the importance of external validation, high quality reporting guidelines and standardisation protocols to ensure generalisability, replication and ultimately clinical implementation.

Predicting radiotherapy-induced xerostomia in head and neck cancer patients using day-to-day kinetics of radiomics features

Background and purpose

The images acquired during radiotherapy for image-guidance purposes could be used to monitor patient-specific response to irradiation and improve treatment personalisation. We investigated whether the kinetics of radiomics features from daily mega-voltage CT image-guidance scans (MVCT) improve prediction of moderate-to-severe xerostomia compared to dose/volume parameters in radiotherapy of head-and-neck cancer (HNC).

Materials and Methods

All included HNC patients (N = 117) received 30 or more fractions of radiotherapy with daily MVCTs. Radiomics features were calculated on the contra-lateral parotid glands of daily MVCTs. Their variations over time after each complete week of treatment were used to predict moderate-to-severe xerostomia (CTCAEv4.03 grade ≥ 2) at 6, 12 and 24 months post-radiotherapy. After dimensionality reduction, backward/forward selection was used to generate combinations of predictors.Three types of logistic regression model were generated for each follow-up time: 1) a pre-treatment reference model using dose/volume parameters, 2) a combination of dose/volume and radiomics-based predictors, and 3) radiomics-based predictors. The models were internally validated by cross-validation and bootstrapping and their performance evaluated using Area Under the Curve (AUC) on separate training and testing sets.

Results

Moderate-to-severe xerostomia was reported by 46 %, 33 % and 26 % of the patients at 6, 12 and 24 months respectively. The selected models using radiomics-based features extracted at or before mid-treatment outperformed the dose-based models with an AUCtrain/AUCtest of 0.70/0.69, 0.76/0.74, 0.86/0.86 at 6, 12 and 24 months, respectively.

Conclusion

Our results suggest that radiomics features calculated on MVCTs from the first half of the radiotherapy course improve prediction of moderate-to-severe xerostomia in HNC patients compared to a dose-based pre-treatment model.

Multimodal image synthesis based on disentanglement representations of anatomical and modality specific features, learned using uncooperative relativistic GAN

Growing number of methods for attenuation-coefficient map estimation from magnetic resonance (MR) images have recently been proposed because of the increasing interest in MR-guided radiotherapy and the introduction of positron emission tomography (PET) MR hybrid systems. We propose a deep-network ensemble incorporating stochastic-binary-anatomical encoders and imaging-modality variational autoencoders, to disentangle image-latent spaces into a space of modality-invariant anatomical features and spaces of modality attributes. The ensemble integrates modality-modulated decoders to normalize features and image intensities based on imaging modality. Besides promoting disentanglement, the architecture fosters uncooperative learning, offering ability to maintain anatomical structure in a cross-modality reconstruction. Introduction of a modality-invariant structural consistency constraint further enforces faithful embedding of anatomy. To improve training stability and fidelity of synthesized modalities, the ensemble is trained in a relativistic generative adversarial framework incorporating multiscale discriminators. Analyses of priors and network architectures as well as performance validation were performed on computed tomography (CT) and MR pelvis datasets. The proposed method demonstrated robustness against intensity inhomogeneity, improved tissue-class differentiation, and offered synthetic CT in Hounsfield units with intensities consistent and smooth across slices compared to the state-of-the-art approaches, offering median normalized mutual information of 1.28, normalized cross correlation of 0.97, and gradient cross correlation of 0.59 over 324 images.

Assessing the Accuracy of Bioluminescence Image-Guided Stereotactic Body Radiation Therapy of Orthotopic Pancreatic Tumors Using a Small Animal Irradiator

Stereotactic body radiation therapy (SBRT) has shown promising results in the treatment of pancreatic cancer and other solid tumors. However, wide adoption of SBRT remains limited largely due to uncertainty about the treatment's optimal fractionation schedules to elicit maximal tumor response while limiting the dose to adjacent structures. A small animal irradiator in combination with a clinically relevant oncological animal model could address these questions. Accurate delivery of X rays to animal tumors may be hampered by suboptimal image-guided targeting of the X-ray beam in vivo. Integration of bioluminescence imaging (BLI) into small animal irradiators in addition to standard cone-beam computed tomography (CBCT) imaging improves target identification and high-precision therapy delivery to deep tumors with poor soft tissue contrast, such as pancreatic tumors. Using bioluminescent BxPC3 pancreatic adenocarcinoma human cells grown orthotopically in mice, we examined the performance of a small animal irradiator equipped with both CBCT and BLI in delivering targeted, hypo-fractionated, multi-beam SBRT. Its targeting accuracy was compared with magnetic resonance imaging (MRI)-guided targeting based on co-registration between CBCT and corresponding sequential magnetic resonance scans, which offer greater soft tissue contrast compared with CT alone. Evaluation of our platform's BLI-guided targeting accuracy was performed by quantifying in vivo changes in bioluminescence signal after treatment as well as staining of ex vivo tissues with γH2AX, Ki67, TUNEL, CD31 and CD11b to assess SBRT treatment effects. Using our platform, we found that BLI-guided SBRT enabled more accurate delivery of X rays to the tumor resulting in greater cancer cell DNA damage and proliferation inhibition compared with MRI-guided SBRT. Furthermore, BLI-guided SBRT allowed higher animal throughput and was more cost effective to use in the preclinical setting than MRI-guided SBRT. Taken together, our preclinical platform could be employed in translational research of SBRT of pancreatic cancer.

Population-based comparative survival analysis of surgery with or without adjuvant radiotherapy and non-operative primary radiotherapy in patients with early-stage oral tongue squamous cell carcinoma

Purpose

Although recent clinical guidelines do allow primary radiotherapy for selected patients with early-stage oral tongue cancer, there has been little knowledge on the treatment outcomes of non-operative radiotherapy using modern treatment techniques. This study evaluated recent prognostic differences between primary radiotherapy and surgical resection in T1‒2N0 oral tongue squamous cell carcinoma.

Methods

Patients diagnosed with T1‒2N0 oral tongue squamous cell carcinoma were identified from the Surveillance, Epidemiology, and End Results database. After propensity score matching, the disease-specific survival of primary radiotherapy and surgery was compared.

Results

From a total of 8,458 patients initially identified, we defined matched cohorts: cohort A, comparing surgery alone vs. primary radiotherapy (n = 230 vs. 230), and cohort B, comparing surgery plus adjuvant radiotherapy vs. primary radiotherapy (n = 230 vs. 230). The 7-year disease-specific survival rates were 77% vs. 35% (cohort A) and 65% vs. 35% (cohort B) (P < 0.001 for all comparisons). Primary radiotherapy was independently associated with worse disease-specific survival in both cohorts A (hazard ratio 4.06; 95% confidence interval 2.53‒6.52) and B (hazard ratio 2.81; 95% confidence interval 1.96‒4.04). Time-course hazard rate function plots showed a distinct short-term risk increment in disease-specific mortality in the primary radiotherapy group.

Conclusion

In the contemporary treatment era, the use of radiotherapy as a definitive treatment resulted in an inferior prognosis in patients with T1‒2N0 oral tongue squamous cell carcinoma. The present population-based data suggest that primary radiotherapy cannot be used as an alternative to surgical management and it needs to be avoided as much as possible in early-stage tumors.

Delivery of online adaptive magnetic resonance guided radiotherapy based on isodose boundaries

Magnetic Resonance-guided Radiotherapy (MRgRT) allows direct monitoring of treated volumes. The aim of this study was to investigate the feasibility of a new gating strategy consisting in using an isodose as boundary. Forty-four patients treated for thoracic and abdominal lesions using MRgRT were enrolled. The accuracy of the new strategy was compared to the conventional one in terms of area improvement available for gating without compromising target coverage. A mean increase of 24% for lung, 15% for liver and 11% for pancreas was observed, demonstrating how the new method can be useful in challenging situations with low dose conformality.