Robust extraction of biological information from diffusion-weighted magnetic resonance imaging during radiotherapy using semi-automatic delineation

A robust semi-automatic delineation workflow using denoised diffusion weighted magnetic resonance imaging for response assessment of patients with esophageal cancer treated with neoadjuvant chemoradiotherapy

Background and Purpose

Diffusion weighted magnetic resonance imaging (DW-MRI) can be prognostic for response to neoadjuvant chemotherapy (nCRT) in patients with esophageal cancer. However, manual tumor delineation is labor intensive and subjective. Furthermore, noise in DW-MRI images will propagate into the corresponding apparent diffusion coefficient (ADC) signal. In this study a workflow is investigated that combines a denoising algorithm with semi-automatic segmentation for quantifying ADC changes.

Materials and Methods

Twenty patients with esophageal cancer who underwent nCRT before esophagectomy were included. One baseline and five weekly DW-MRI scans were acquired for every patient during nCRT. A self-supervised learning denoising algorithm, Patch2Self, was used to denoise the DWI-MRI images. A semi-automatic delineation workflow (SADW) was next developed and compared with a manually adjusted workflow (MAW). The agreement between workflows was determined using the Dice coefficients and Brand Altman plots. The prognostic value of ADCmean increases (%/week) for pathologic complete response (pCR) was assessed using c-statistics.

Results

The median Dice coefficient between the SADW and MAW was 0.64 (interquartile range 0.20). For the MAW, the c-statistic for predicting pCR was 0.80 (95% confidence interval (CI):0.56-1.00). The SADW showed a c-statistic of 0.84 (95%CI:0.63-1.00) after denoising. No statistically significant differences in c-statistics were observed between the workflows or after applying denoising.

Conclusions

The SADW resulted in non-inferior prognostic value for pCR compared to the more laborious MAW, allowing broad scale applications. The effect of denoising on the prognostic value for pCR needs to be investigated in larger cohorts.

Recommendations for improved reproducibility of ADC derivation on behalf of the Elekta MRI-linac consortium image analysis working group

BACKGROUND AND PURPOSE

The apparent diffusion coefficient (ADC), a potential imaging biomarker for radiotherapy response, needs to be reproducible before translation into clinical use. The aim of this study was to evaluate the multi-centre delineation- and calculation-related ADC variation and give recommendations to minimize it.

MATERIALS AND METHODS

Nine centres received identical diffusion-weighted and anatomical magnetic resonance images of different cancerous tumours (adrenal gland, pelvic oligo metastasis, pancreas, and prostate). All centres delineated the gross tumour volume (GTV), clinical target volume (CTV), and viable tumour volume (VTV), and calculated ADCs using both their local calculation methods and each of the following calculation conditions: b-values 0-500 vs. 150-500 s/mm2, region-of-interest (ROI)-based vs. voxel-based calculation, and mean vs. median. ADC variation was assessed using the mean coefficient of variation across delineations (CVD) and calculation methods (CVC). Absolute ADC differences between calculation conditions were evaluated using Friedman's test. Recommendations for ADC calculation were formulated based on observations and discussions within the Elekta MRI-linac consortium image analysis working group.

RESULTS

The median (range) CVD and CVC were 0.06 (0.02-0.32) and 0.17 (0.08-0.26), respectively. The ADC estimates differed 18% between b-value sets and 4% between ROI/voxel-based calculation (p-values < 0.01). No significant difference was observed between mean and median (p = 0.64). Aligning calculation conditions between centres reduced CVC to 0.04 (0.01-0.16). CVD was comparable between ROI types.

CONCLUSION

Overall, calculation methods had a larger impact on ADC reproducibility compared to delineation. Based on the results, significant sources of variation were identified, which should be considered when initiating new studies, in particular multi-centre investigations.

Semi-automatic tumor segmentation of rectal cancer based on functional magnetic resonance imaging

•

Machine learning on magnetic resonance images (MRI) was used for tumor segmentation.

•

Voxelwise machine learning with morphological post-processing achieved good segmentation results.

•

Combining T2-weighted with functional MRI improved semi-automatic tumor segmentation.

•

Dynamic contrast enhanced MRI was the most valuable functional MRI information.

•

Tumor volume and interobserver variation were linked to measured segmentation quality.

Background and purpose

Tumor delineation is required both for radiotherapy planning and quantitative imaging biomarker purposes. It is a manual, time- and labor-intensive process prone to inter- and intraobserver variations. Semi or fully automatic segmentation could provide better efficiency and consistency. This study aimed to investigate the influence of including and combining functional with anatomical magnetic resonance imaging (MRI) sequences on the quality of automatic segmentations.

Materials and methods

T2-weighted (T2w), diffusion weighted, multi-echo T2*-weighted, and contrast enhanced dynamic multi-echo (DME) MR images of eighty-one patients with rectal cancer were used in the analysis. Four classical machine learning algorithms; adaptive boosting (ADA), linear and quadratic discriminant analysis and support vector machines, were trained for automatic segmentation of tumor and normal tissue using different combinations of the MR images as input, followed by semi-automatic morphological post-processing. Manual delineations from two experts served as ground truth. The Sørensen-Dice similarity coefficient (DICE) and mean symmetric surface distance (MSD) were used as performance metric in leave-one-out cross validation.

Results

Using T2w images alone, ADA outperformed the other algorithms, yielding a median per patient DICE of 0.67 and MSD of 3.6 mm. The performance improved when functional images were added and was highest for models based on either T2w and DME images (DICE: 0.72, MSD: 2.7 mm) or all four MRI sequences (DICE: 0.72, MSD: 2.5 mm).

Conclusion

Machine learning models using functional MRI, in particular DME, have the potential to improve automatic segmentation of rectal cancer relative to models using T2w MRI alone.