Keywords:
Gastrointestinal tract, Oncology, MR, MR-Diffusion/Perfusion, Outcomes analysis, Diagnostic procedure, Cancer
Authors:
L. Yang, B. Wu; Chengdu/CN
DOI:
10.1594/ecr2018/C-0510
Methods and materials
Patient population: In this retrospective study 43 consecutive patients (ages 31-76,
mean age 55) with LARC (cT3 or T4,
N+) who have completed neoadjuvant chemoradiotherapy (NCRT) and surgery were included between October 2015 and June 2017.
The criteria of pTRG were: 0 (pCR,
no remaining viable cancer cells); 1 (only small clusters cancer cells remaining); 2 (residual cancer remaining,
but with predominant fibrosis); 3(minimal or no tumor kill,
or extensive residual cancer).
Among the enrolled patients,
17 were pTRG0,
7 were pTRG1,
11 were pTRG2,
and 8 were pTRG3.
Exclusion criteria were: (a) lack of complete baseline and preoperational MRI images; (b) obvious artifacts interference; (c) time interval between preoperational MRI and surgery larger than one month; (d) patients without pathological results.
Data acquisition: All MRI examinations were performed on a 3 Tesla (T) imaging unit,
by using an eighteen-element pelvic phased-array body coil.
Patients were placed in supine position,
without specific bowel preparation.
The examination protocols did not change over the study period.
The MRI protocols consisted of high resolution T2 weighted fast spin echo (T2W-FSE) sagittal,
oblique axial and oblique coronal plane imaging (perpendicular and parallel to the longest tumor axis as identified on the sagittal images) with a slice thickness of 3mm,
and rs-EPI DWI imaging[7] with b values of 0 and 1000 s/mm2.
Imaging evaluation:mrTRG was assessed based on similar principles to pTRG described above.
Both baseline and preoperational T2-weighted imagies and diffusion weighted imagies were reviewed by two radiologists to reach a consensus of mrTRG,
blinded to clinical and pathological data.
Statistical analysis: Patients were divided into complete regression group (CR,
pTRG0 and mrTRG0) and non-CR group (pTRG1-3 and mrTRG1-3),
as well as good response group (GR,
pTRG0-1 and mrTRG0-1) and poor response group (PR,
pTRG2-3 and mrTRG2-3) respectively.
The kappa-test was used to evaluate the overall correspondence of mrTRG and pTRG,
so as to each binary group.
The Chi-square test was done to identify the discriminative power of mrTRG in selecting pCR and GR.