Comparison of the diagnostic performance of changes in signal intensity and volume from multiparametric MRI for assessing response of rectal cancer to neoadjuvant chemoradiotherapy

To evaluate the change in signal intensity (SI) and volume (V) from multiparametric magnetic resonance imaging (MRI) for assessing the response of locally advanced rectal cancer (LARC) to chemoradiotherapy (CRT).


INTRODUCTION
The use of multiple treatment modalities for locally advanced rectal cancer(LARC) has markedly improved patient outcomes during the last three decades because of the use of chemoradiotherapy (CRT). 1,2 CRT provides major benefits in terms of tumor regression grade (TRG), a factor that is significantly and independently associated with improved survival and disease-free survival (DFS). 3 For example, Huh et al. compared the prognostic significance of TRG in patients with rectal cancer after preoperative CRT and found better 5-year overall survival (OS) and DFS in patients with in those with TRG-0 (complete response) than in those with TRG-1 (moderate), or TRG-2-3(minimal-poor). 4 Suzuki et al. used a modified classification system based on pathological T (ypT) stage and TRG, and showed that a less advanced stage was independently related to improved DFS. 5 Although the advantage of establishing the preoperative TRG is not clear because this information is provided by the pathology results, it may allow more individualized treatment, such as a non-surgical "watch and wait" approach, because approximately 10 to 30% of rectal cancer patients achieve pCR after CRT. 1,2,[6][7][8] Magnetic resonance imaging (MRI) is now the gold standard for initial staging of rectal cancer and assessment of restaging after neoadjuvant CRT. 6,7,9,10 Several studies reported the use of TRG based on MRI (mrTRG) following preoperative CRT for predicting pathological TRG (pTRG), DFS, and OS in patients with LARC. 1,[11][12][13][14][15] However, there are differences between mrTRG and pTRG, and there is no consensus on whether mrTRG can be used as a surrogate of pTRG 6,12,16 because it is difficult to distinguish tumors from tissue fibrosis based on MRI signal intensity (SI) after CRT. 3,17 Consequently, some studies performed quantitative analyses of MRI results with morphological evaluations to assess TRG after CRT based on tumor SI and volume (V), and showed promising results. [16][17][18][19][20][21][22] Other studies reported that tumor SI and V had some potential additional diagnostic value for patients who achieved good pathological response. 2,16,19,23 Previous studies examined the use tumor SI and V to discriminate pCR or other responsive groups from non-pCR or other nonresponsive groups. 19,24 However, no study has yet reported the use of quantitative multiparametric MRI to determine different TRGs in LARC. Therefore, the purpose of this study of patients with LARC was to evaluate the change in relative SI (SIR) and change of V from multiparametric MRI before and after CRT to identify patients who may benefit from a "wait-and-see" strategy (complete responders) from other patients who should undergo surgery (partial responders, poor responders, non-responders).

This retrospective study was approved by the Institutional Review
Board and Research Ethics Committee of our Hospital. The need for written informed consent was waived.

Patients
Clinical data were obtained from the hospital database for 168 consecutive patients with rectal cancer who received total mesorectal resection (TME) after CRT between October 2017 and October 2019 ( Figure 1). 2,24

MRI protocol
Patients were examined using different MRI machines (Table 1)

Evaluation of SI and V
Two gastrointestinal radiologists with expertise in rectal cancer diagnosis (7 years for one and 5 years for the other) who were blinded to patient clinical data independently calculated tumor V and SI by manually tracing the tumor boundaries on the axial images and placing free-hand regions-of-interest (ROIs), which provided the area of the lesion for each tumor-containing section ( Figure 2  In some patients, low SI zones were not identified on post-CRT ADC images, in which case the ROIs were positioned at the location of the tumor bed before CRT ( Figure 3). The SI of the iliopsoas muscle (SIm) was used as reference tissue, with careful avoidance of any intramuscular fat. 16,[25][26][27] The absolute change in V and SI values between pre-CRT and post-CRT measurements was expressed as Δ, and the percentage change was expressed as %Δ. 19,25,28 A third independent radiologist, who had 10 years of expertise in gastrointestinal diagnostics, independently assessed each pretreatment MRI T staging. This radiologist was also blinded to the clinical, histopathological, and other characteristics of the patients.

Pathological evaluation after CRT
Surgically resected specimens were pathologically analyzed according to the seventh edition of the American Joint Committee on Cancer (AJCC) TNM staging system and the Ryan et al. staging system with the 4-point TRG scoring system. [29][30][31] The TRG assessment of the pathology specimen was only for the primary tumor, with no consideration of regional LN status or other loco-regional or residual localization.

Statistical analysis
Categorical values were expressed as numbers and percentages and continuous variables as means ± standard deviations (normal distribution) or medians and interquartile ranges (non-normal distribution).
The Kruskal-Wallis test (continuous variables) and the χ 2 test (categorical variables) were used to assess the significance of differences.

RESULTS
We enrolled 82 patients with LARC who fulfilled the inclusion and exclusion criteria (Figure 1). The average age was 51.67 ± 10.16 years and there were 61 men (74.4%) and 21 women (25.6%).

Diagnostic performance for TRG
We performed ROC analysis to compare the diagnostic performance   Figure 4).

Interobserver agreement
Our
Although our ROC analysis indicated no significant differences among these AUC values, ADC-%△V * T2W-%△SIR had the highest AUC values. To the best of our knowledge, no prior studies comprehensively compared the diagnostic performances of change in V and SIR using T2W, ADC, and ceT1W MRI sequences to predict different TRG in patients with LARC after CRT.
Some studies showed that MR quantification is useful for evaluating cancer response after CRT. 33 Although our results are encouraging, there were some limitations to our study. First, this was a retrospective study and there were small numbers of patients in the different TRG groups. Second, our data were from multiple MRI machines because it was collected retrospectively.
However, except for the fair inter-observer correlation of TRG with ADC-%△SIR, other MRI measurements had excellent inter-observer correlation, consistent with the results of prior studies. 2,19,21,[25][26][27][28]38,39 Third, the selection of an ROI and comparison of MRI results before and after CRT was somewhat subjective; however, to overcome these limitations, we evaluated a relatively large area of cancerous tissue. 37 Finally, there were variations in the timing between the pre-and post-CRT MRI examinations (as indicated by the large standard deviation), and there was no uniform and standard CRT scheme. These factors could have increased the risk of selection bias. However, we used the change in volume and SIR, and the degree of tumor change after CRT was incorporated into the corresponding TRG groups, and this did not influence grouping of cases by TRG. 18,19 In conclusion, despite the several limitations of this study, %△V and %△SIR on T2W, ADC and ceT1Wprovided good to excellent predictions of TRG after CRT in patients with LARC. These results require confirmation by a large prospective cohort study. The advantage of establishing the preoperative TRG in these patients is that it allows a more effective management that can be specifically tailored to individual patients.

AUTHOR CONTRIBUTIONS
Xin Li designed the study; Xin Li, Zhengwu Tan, Lingling Xie, and Zhenyu Lin collected the data; Lan Cheng collected MRI scanning parameters; Xin Li, Zhengwu Tan, and Lan Zhang reviewed the imaging data; Zhengwu Tan and Lan Zhang measured and analyzed the data, reviewed the charts, and interpreted the data; Xin Li and Zhengwu Tan wrote the manuscript; Ping Han modified the manuscript. All authors have read and approved the manuscript.