SS 201b - Rectal cancer: response assessment and diagnostic biomarkers
Performance of high-resolution MRI in predicting tumoural response to neoadjuvant chemoradiotherapy in patients with rectal cancer
Purpose: To determine the performance of high-resolution MRI parameters in predicting complete tumoral response in patients with locally advanced rectal cancer treated with neoadjuvant chemoradiotherapy (CRT).
Methods and Materials: Population consisted of 52 patients with rectal cancer, who underwent pre-treatment 1.5-T MRI. Tumour ADC (DWI) measurements (10-3 mm2/s), lesion site, tumour longitudinal extent, distance between lesion and puborectalis muscle, levator ani muscles infiltration, depth of extramural spread and lymph nodes involvement were appreciated. After CRT, all patients underwent complete surgical resection and the surgical specimen served as gold standard. Area under receiver operating characteristic curve (AUROC) analysis was performed to assess the discriminatory power of MRI parameters to predict complete response.
Results: Pathological complete response, partial response and no response were found in 6, 22 and 23 patients, respectively. Among analysed parameters mean ADC values, depth of extramural spread and lymph nodes involvement were independent predictors of pathological complete response (p<0.0001). Independently, pathological complete response can be predicted by mean ADC values (AUROC=0.90, Se=93.8%, Sp=73.3%), depth of extramural spread (AUROC=0.72, Se=36.73%, Sp=97.78%) and absence of lymph nodes involvement (AUROC=0.87, Se=80.82%, Sp=93.33%), respectively. Together, the three parameters can accurately assess pathological complete response with an AUROC of 0.92 (Se=79.59%, Sp=93.33%).
Conclusion: Baseline high-resolution MRI parameters have the potential to act as imaging biomarkers of tumoral response to neoadjuvant chemoradiotherapy.
Purpose: A complete response to CRT may candidate patients with local advanced rectal cancer (LARC) to a conservative management. Hence, the non invasive prediction of response to CRT is of pivotal importance. Aim was to assess the value of DWI in the prediction of pathological response to CRT in LARC.
Methods and Materials: 29 pts with LARC underwent 1.5T MRI before the beginning of CRT (preMRI), after 6 week of CRT (midMRI), and at the end of CRT (postMRI). High-resolution axial-T2w sequences on the three planes and DWI sequences (b:0-200-600-1000) were acquired. Cancer volume was segmented on T2w (VT2) and b:1000 (Vb1000) images, and the percentages of modification over time (ΔV) were evaluated. VT2 was coregistered on ADC map, and a volumetric histogram-analysis of ADC performed. After surgery patients were classified according to Tumour Regression Grade (TRG) in Non-Responder (NR) when TRG=0-2, Partial Responder (PR) when TRG=3, and Complete Responder (CR) when TRG=4.
Results: At histopathology 8 patients were CR, 13 PR, 8 NR. A good correlation between VT2 and Vb1000 was observed at each time point (p<0.01), with lower Vb1000values than VT2 (p<0.01). At multivariate analysis ΔV T2 >81.5% at midMRI, Vb1000<3 cc at postMRI and mean ADC>1,4 at preMRI resulted able to predict CR with an accuracy of 86%, 89% and 76%, respectively. At preMRI ADC was higher in CR as well as ADC skewness (p<0.01).
Conclusion: DWI may improve the MRI capability to predict the LARC response to CRT; in particular, Vb1000 at postMRI increases its accuracy.
The value of ADC measurements for assesing treatment response of neoadjuvant chemoradiotherapy in patients with localy advanced rectal cancer
Purpose: To evaluate the diagnosis performance of ADC measurements in assessing treatment response to neoadjuvant chemoradiation therapy (CRT) in patients with locally advanced rectal cancer.
Methods and Materials: Forty-eight consecutive patients with pathologically confirmed rectal adenocarcinoma, who had undergone pretreatment MRI (1.5T; T2TSE/DWI) were retrospectively collected. ADC measurements based on variable b-value (50,400,800 x 10-3 mm2/s) were performed using a variable ROI size and six diagnostic measurement tools were computed: min ADC, max ADC, mean ADC and their corresponding ratio with ADC values of normal rectal wall. Patients were classified as responders and non-responders according to rectal cancer regression grade. Univariate and multivariate logistic regression models were performed to predict complete response to treatment and the diagnostic performance was assessed using the area under the receiver operating characteristic curve analysis (AUROC).
Results: Mean ADC values presented significantly lower values in patients with complete response compared to those with partial or no response (692.16±43.7 vs 835.17±88.38, p<0.001) and had the best diagnostic performance (AUROC=0.91, Se=100%, Sp=82.50%, cut-off 0.758 × 10(-3) mm(2)/s) in predicting complete response after CRT.
Conclusion: Mean ADC values perform best in predicting complete response after CRT in patients with locally advanced rectal cancer.
Locally advanced rectal cancer: predicting response to neoadjuvant chemoradiotherapy using apparent diffusion coefficient textures
Purpose: To evaluate whether ADC textures could predict patient with locally advanced rectal cancer (LARC) who would not response to neoadjuvant chemoradiotherapy (NCRT).
Methods and Materials: Thirty-two patients who underwent MRI including diffusion-weighted imaging at a 3.0 T system before NCRT were enrolled. Texture analysis of pre-therapy ADC mapping was performed, and a total of 135 ADC textures as well as routine mean ADC value of the primary tumour were extracted for each patient. Texture parameters and mean ADC value were compared between responsive group and non-responsive group. Receiver operating characteristic (ROC) curves were performed to evaluate the predictive performance of each significant parameter.
Results: Among the texture parameters, skewness, energy, Gabor-43, high-intensity emphasis, and low-intensity small area emphasis significantly differed between responsive and non-responsive groups (0.72 vs 0.30, p = 0.020; 0.11 vs 0.08, p = 0.045; 3.34 vs 2.19, p = 0.028; 65.67 vs 98.06, p = 0.045; 21.49 vs 8.06, p = 0.038). Further, skewness, high-intensity emphasis and low-intensity small area emphasis were identified as independent predictors for non-responders to NCRT, with areas under the ROC curves (AUC) of 0.754, 0.729, and 0.739, respectively.
Conclusion: Several texture features derived from pre-therapy ADC image could potentially be helpful to predict patients with LARC who would not response to NCRT.
DCE-MRI shows slower flow and more homogeneous vascularity in responding tumours after CRT for rectal cancer
Purpose: To evaluate whether pre-chemoradiation(CRT) DCE-MRI can predict response to CRT in rectal cancer.
Methods and Materials: 20 patients with locally advanced rectal cancer underwent DCE-MRI with the contrast agent gadofosveset-trisodium before CRT. DCE-MRI was processed with the Intellispace Discovery research platform(Philips Healthcare). One reader delineated whole-tumour volumes on DWI. Semiquantitative DCE-parameters based on the enhancement curve(AUC(60), (AUC of)time to peak(TTP), maximum enhancement, bolus arrival time(BAT), initial signal excess (ISE), mean transit time(MTT), wash-in- and wash-out-parameters) were compared between patients with complete(CR;ypT0) or good response(GR;ypT0-2) and non-responders. Additionally, heterogeneity of DCE parameters (measured by coefficient of variance(CoV:sd/mean)) was compared.
Results: 8/20(40%) had CR. Almost all DCE-parameters were lower in CR and GR (p>0.05). MTT and BAT were higher in CR, with BAT significantly higher(p=0.015). In good responders the BAT was also significantly higher(p=0.028). Heterogeneity was higher in CR for almost all DCE-parameters, except for BAT and wash-out, that had somewhat lower heterogeneity (p>0.05). The higher heterogeneity of AUCTTP, wash-in and AUC60 in CR showed a trend towards statistical significance (p=0.062-0.099). In good responders AUC60 and TTP were significantly more heterogeneous than in poor responders(p=0.024-0.041). Washout slope(>0.05) and BAT (p=0.012) were less heterogeneous in good responders.
Conclusion: Bolus arrival time is longer and less heterogeneous in CR and GR after CRT, potentially reflecting slower but more homogeneous inflow of the contrast bolus. This could represent prolonged and more homogeneous exposure to chemotherapy and reduced hypoxia, leading to more response. Most other DCE parameters were lower and more heterogeneous in responders.
A modified 3-point MRI-predicted tumour regression grade incorporating diffusion-weighted image: locally advanced rectal cancer
Purpose: To evaluate the prognostic relevance of a modified 3-point MRI-predicted tumour regression grade (mrTRG) incorporating diffusion-weighted image (DWI) in patients with locally advanced rectal cancer (LARC) after preoperative chemoradiotherapy (PCRT).
Methods and Materials: Between March 2012 and September 2013, 118 consecutive patients with mid/lower LARC, who underwent PCRT followed by surgery, were enrolled in this retrospective study. Two radiologists in consensus assessed the mrTRG based on T2-weighted images and high b-value DWIs (0, 1000 s/mm2) using the following grade: Grade 0, good regression (no obvious tumour); Grade 1, intermediate regression (dominant fibrosis; regression > 50%); Grade 2, poor regression (dominant tumour; regression ≤ 50%). A multivariate Cox proportional-hazard regression was performed to evaluate the association of the modified mrTRG with the 3-year disease-free survival (DFS). A Kaplan-Meier analysis with a log-rank test was used to compare the DFS rate between responder (Grade 0 and 1) and non-responder (Grade 2) groups.
Results: The modified mrTRG (adjusted HR, 2.505; 95% CI 1.231 - 5.100) was independently associated with the 3-year DFS (P = 0.011). Additionally, there was a significant difference in the 3-year DFS rate between responder (73.8%; 95% CI 64.2 - 81.3%) and non-responder (41.7%; 95% CI 10.9 - 70.8%) groups (P = 0.028).
Conclusion: In patients with mid/lower LARC, the modified 3-point mrTRG incorporating DWI was independently associated with the 3-year DFS after PCRT followed by surgery. The grading scale may be used as a surrogate for expected prognosis to PCRT prior to surgery. Further prospective trials are warranted.
Purpose: MRI tumour volumetry (TV) is a recently described imaging concept that may have advantages over radiological TNM staging of rectal cancer in assessing tumour response to chemoradiotherapy. This study aims to investigate the performance of TV against traditional radiological and pathological staging modalities, and introduces the novel concept of tumour volumetry to mesorectal volume ratio (TMVR).
Methods and Materials: Patients with histologically proven rectal cancer treated at a single cancer institute by surgery alone were selected from a prospective database. Axial MRI images for T2 to T4 disease were selected and transferred to an offline workstation (Prosoma ®). A single investigator delineated each axial slice for the rectal tumour and mesorectum defining the tumour (TV) and mesorectal volumes. The ratio of tumour volume to mesorectal volume (TMVR) was calculated. Correlations between the tumour volume, TMRV and tumour length were compared with MRI and pathological T staging
Results: 35 patients with rectal cancer treated without neoadjuvant therapy were analysed. A distinct correlation was observed between increasing TV and advancing ordinal T stage ( p=0.027 , Logistic regression). TV correlated better with pathological T stage than radiological T stage. In particular there was improved prediction of pathological T3 subset (a-d) with volumetric over MR-T stage. Correction of TV with mesorectal volume did not improve accuracy.
Conclusion: Tumour volumetry is a promising predictor of final pathological T stage. This reflects that volumetry is a scale which potentially smooths out differences between discrete T staging categories.
Accuracy of MRI for predicting anterior peritoneal reflection involvement for locally advanced rectal cancer: prospective comparison with operation findings
Purpose: To assess the diagnostic accuracy of preoperative rectal MRI for rectal cancer involvement through anterior peritoneal reflection (PR) in comparison with operation findings.
Methods and Materials: This prospective study was approved by Institutional review board; informed consent was waived. Eighty-three consecutive patients (Group 1: direct surgery, n=56; Group 2: neoadjuvant CCRT and surgery, n=27) with locally advanced rectal cancer were enrolled in this prospective study. All patients underwent rectal MR imaging using 3T systems. The identification of anterior PR was assessed using a 5-point scale. The possibility of anterior PR involvement was designated as positive or negative and all radiologic assessment were obtained in independent review by two radiologists. Diagnostic accuracy of MRI was obtained by operation findings and surgical record.
Results: More than 90% patients showed relatively good anterior PR identification regardless of neoadjuvant CCRT (Chi square test, P = .580) and there was no statistical difference in two reviewers (McNemar test, P = .443). The fifteen patients (15 of 83 patients, 18.1%) were clinically suspected anterior PR involvement during surgery. The sensitivity, specificity, positive predictive value, and negative predictive value of anterior PR involvement in the preoperative MRI were 69.2%, 78.3%, 47.4%, and 90%, respectively. The diagnostic accuracy of MRI for predicting for anterior peritoneal reflection involvement was 76.3%.
Conclusion: MR assessment of tumour involvement through anterior peritoneal reflection has relatively low sensitivity and positive predictive value, although preoperative rectal MR provides accurate anatomical information regarding anterior PR with high conspicuity.
Purpose: It is believed that chemoradiation for low rectal cancer increases sphincter preservation. Aim was to evaluate with MRI whether sphincter preservation is increased by chemoradiation and whether MRI can predict sphincter preservation after chemoradiation.
Methods and Materials: A radiologist independently evaluated the T2-weighted MRIs (in 3 directions) in 47 patients before and after CRT with tumours <5 cm from the anorectal junction (ARJ) and measured distance of the lower tumour pole to the ARJ. Also, a confidence level score for feasibility of sphincter preservation was scored (CL=0 definitely no sphincter preservation, CL4=sphincter preservation definitely possible). Likelihood for sphincter preservation before and after CRT was compared and receiver operator characteristics(ROC) curves with area under the curve (AUC) were calculated.
Results: Mean distance from ARJ increased significantly during CRT from 21±16mm pre-CRT to and 31±18mm post-CRT (P<0.001). In 42% sphincter preservation was deemed not feasible pre-CRT, which decreased to 23% after CRT. AUC for sphincter preservation based on confidence level score was 0.84(0.72-0.96), with sensitivity of 100% and specificity of 44%. Based on post-CRT height measurement AUC was 0.87(0.76-0.98), with optimal size cut-off at 26 mm (sens: 86%, spec: 71%).
Conclusion: This is the first study to show that CRT increases the distance to the ARJ and thus leads to a higher rate of sphincter preservation. MRI can accurately predict sphincter preservation after CRT.
Organ preservation for clinical complete responders after chemoradiation for rectal cancer, does timing of selection matter?
Purpose: Wait-and-see policy could be offered to clinical complete responders(cCR) after neoadjuvant chemoradiation(CRT). In patients with near cCR a second restaging can be considered to evaluate whether response becomes complete. Aim was to evaluate whether timing of inclusion influences outcome, by comparing patients included for wait-and-see at initial assessment with patients who were included after a second assessment 3 months later.
Methods and Materials: (Near) CRs underwent endoscopy+MRI+DWI ±8 weeks post-CRT. 103 had typical cCR and were selected for wait-and-see(W&S-1). The other 67 patients had near cCR, 19 underwent TEM and the 48 had a second assessment after 3 months, after which 43 were included for wait-and-see(W&S-2); 5 had TME. 3-6 monthly follow-up(MRI+DWI+endoscopy) was performed. 2-year local regrowth(LR), disease-free survival(DFS) and overall survival(OS) were compared between W&S-1 and W&S2. Multivariable Cox regression analyses were performed to assess whether inclusion timing was predictive for higher LR.
Results: 2-year LR in W&S-1 was 18% and 23% in W&S-2. 2-year DFS was 82% in W&S-1 and 78% in W&S-2. 2-year OS was 99% in W&S-1 and 98% in W&S-2 (all: p>0.05). The timing of inclusion was not predictive of more local regrowth or lower DFS. All LRs were detected early, so that they could be easily salvaged with standard TME.
Conclusion: A second response assessment 3 months after the first post-CRT staging can offer wait-and-see to more patients, who would otherwise be referred for TME. However, the 2-year LR is slightly higher in the late inclusion group, but this does not affect OS.
Purpose: Diffusion-weighted imaging (DWI) is increasingly included in standard rectal MRI-protocols. Single-shot echo planar imaging (EPI) is the most commonly used DWI-method. DWI-EPI is prone to susceptibility artefacts, mainly caused by air in the rectal lumen. Aim was to assess whether application of a micro-enema can reduce these artefacts.
Methods and Materials: 50 patients were included who underwent sequential DWI-MRIs (1.5T; highest b-value b1000) during follow-up as part of a wait-and-see approach after chemoradiotherapy. Until March 2014 DWI-MRIs were acquired without bowel preparation, thereafter a micro-enema (Microlax®;5 ml) was routinely given ±15 minutes prior to the DWI-MRI. The presence/severity of air artefacts was scored by 2 readers in consensus on each b1000 DWI-scan with a confidence level, ranging from 0 (no artefact) to 5 (severe artefact). A score >= 3 (moderate-severe) was considered significant. The presence/severity of artefacts was compared between DWI-scans with/without a micro-enema. Potential confounding factors (age, gender, DWI acquisition parameters, MRI-scanner, endoscopy prior to MRI) were taken into account.
Results: In total 335 DWI-MRIs were assessed. Significant air artefacts were seen in 24.3% (no micro-enema) vs 3.7% (with micro-enema). Using binary logistic regression with samples clustered by patient, the odds ratio between the use of a micro-enema and presence of significant artifacts was 0.12 (95% CI 0.04-0.40), p=0.0005. None of the parameters assessed for possible confounding significantly altered this effect.
Conclusion: The use of a micro-enema prior to rectal EPI-DWI examinations significantly reduces the amount and severity of air artefacts, compared to examinations without bowel preparation.