1820

Correlation of MR based textural analysis parameters with MR tumor regression score and its potential role in predicting response to long course chemoradiotherapy in patients with locally advanced rectal cancers [LARC].
Karthik Ganesan1, Ankit Jain1, Shivsamb Jalkote1, Alam Shah1, Slesha Bhalja1, Balaji Ganeshan2, and Swarup Nellore1

1Sir.H.N.Reliance Foundation Hospital & Research Center, Mumbai, India, 2Institute of Nuclear Medicine, University College London, London, United Kingdom

Synopsis

Treatment of locally advanced rectal cancers (cT3/4, cN1/2) with neoadjuvant chemoradiotherapy leads to a reduction in tumor size and enhances the likelihood of tumor resectability and sphincter preservation with low local recurrence rates. In this study we attempted to correlate first order MR textural parameters with the MR tumor regression [TRG] score determined on the post-treatment MRI and to also predict response to long course chemoradiotherapy in locally advanced rectal cancers. We retrospectively included 20 patients with pathologically proven rectal carcinoma who underwent long course chemoradiotherapy [LCCRT] following a pre-treatment MRI. All 20 patients had a post-treatment MRI after a 6-week interval at which MR tumor regression [TRG] scores were determined based on the T2-w and DW-ADC maps. TRG was represented as MR_TRG [TRG 1,2 - Complete Response [CR] = 7; TRG 3 - Partial Response [PR] = 11; TRG 4,5 - Minimal / No response [NR] = 2). Texture analysis was carried out on axial T2-w and ADC images by delineating a 2-D region of interest around the tumour. The results showed that First order MRTA features derived from the post-treatment MR T2-w and ADC images can help differentiate between responders [TRG 1 and 2] and non-responders [TRG 3, 4 and 5] in patients with locally advanced rectal cancers who underwent long course chemoradiotherapy.

Introduction

Treatment of locally advanced rectal cancers (cT3/4, cN1/2) with neoadjuvant chemoradiotherapy leads to a reduction in tumor size and enhances the likelihood of tumor resectability and sphincter preservation with low local recurrence rates. Neoadjuvant CRT followed by total mesorectal excision [TME] is the current standard treatment guideline used worldwide for patients with low rectal cancer. The rationale behind TME has been further validated by the understanding of the importance of the circumferential resection margins and the quality of TME, both prognostic markers of recurrence and survival. In this study we attempted to correlate first order MR textural parameters with the MR tumor regression [TRG] score determined on the post-treatment MRI and to also predict response to long course chemoradiotherapy in locally advanced rectal cancers [1, 2, 3].

Methods

We retrospectively included 20 patients with pathologically proven rectal carcinoma who underwent long course chemoradiotherapy [LCCRT] following a pre-treatment MRI. All 20 patients had a post-treatment MRI after a 6-week interval at which MR tumor regression [TRG] scores were determined based on the T2-w and DW-ADC maps. TRG was represented as MR_TRG [TRG 1,2 - Complete Response [CR] = 7; TRG 3 - Partial Response [PR] = 11; TRG 4,5 - Minimal / No response [NR] = 2). Texture analysis was carried out on axial T2-w and ADC images by delineating a 2-D region of interest around the tumour. MR texture analysis [MRTA] comprised of a filtration-histogram technique using a commercially available research software (TexRAD - Feedback Medical Ltd - www.fbkmed.com, Cambridge, UK). Filtration step extracted and enhanced features of different intensity and sizes corresponding to spatial scale filter (SSF) which varied from 0 (without-filtration), 2mm (fine texture scale), 3-5mm (medium texture scale) and 6mm (coarse texture scale). Quantification of texture using statistical and histogram based analysis comprised of mean intensity, standard-deviation, entropy, mean of positive pixels, skewness and kurtosis. Difference in MRTA for responders and non-responders to LCCRT were assessed using non-parametric Mann Whitney test [MR_TRG_0_CR_1_PRNR (binary - 2 categories where 0 was CR and 1 was PR/NR), and, Kruskal Wallis test [MR_TRG_0_CR_1_PR_2_NR (3 categories where 0 was CR and 1 was PR and 2 was NR)].

Results

On T2-w images without filtration [ssf = 0] the sd difference [p = 0.03], entropy difference [p = 0.046], and entropy percentage change [p = 0.046] and at fine texture [ssf = 2] the mpp difference [p = 0.002] and mpp percentage change [p = 0.024] showed statistically significant differences between responders and partial / non-responders on the Mann Whitney tests. On T2 fine texture [ssf = 2] statistically significant difference in mpp difference [p = 0.014] and mpp percentage change [p = 0.024] was noted between responders and partial / non-responders on the Kruskal Wallis test. ADCs at medium texture [ssf = 4 and 5] the mean percentage change [p = 0.006 - 0.019] and sd difference [0.037] showed statistically significant differences between responders and partial / non-responders on the Mann Whitney tests. On ADCs, statistically significant differences in entropy were noted between responders and partial / non-responders on the Kruskal Wallis test, with the entropy difference ranging between (p = 0.021 [ssf = 0]; 0.027 [ssf = 2]; 0.024 - 0.038 [ssf = 3-5] and 0.024 [ssf = 6) and entropy percentage change (p = 0.027 [ssf = 0]; 0.031 [ssf = 2]; 0.038 - 0.042 [ssf = 3-5] and 0.047 [ssf = 6])


Discussion

First order MRTA features derived from the post-treatment MR T2-w and ADC images can help differentiate between responders [TRG 1 and 2] and non-responders [TRG 3, 4 and 5] in patients with locally advanced rectal cancers who underwent long course chemoradiotherapy.

Conclusion

Post-treatment MR Textural Analysis parameters statistically correlate with the MR Tumor Regression Scores and can potentially help predict response to long course chemoradiotherapy.

Acknowledgements

No acknowledgement found.

References

1.Hocquelet A et al. Pre-treatment magnetic resonance-based texture features as potential imaging biomarkers for predicting event free survival in anal cancer treated by chemoradiotherapy. Eur Radiol. 2018 Jul;28(7):2801-2811.

2. Jalil O et al. Magnetic resonance based texture parameters as potential imaging biomarkers for predicting long-term survival in locally advanced rectal cancer treated by chemoradiotherapy. Colorectal Dis. 2017 Apr;19(4):349-362.

3. De Cecco CN et al. Performance of diffusion-weighted imaging, perfusion imaging, and texture analysis in predicting tumoral response to neoadjuvant chemoradiotherapy in rectal cancer patients studied with 3T MR: initial experience. Abdom Radiol (NY). 2016 Sep;41(9):1728-35.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
1820