SS 216 - Imaging and predicting treatment response and outcome in oncology
SS 216 - Imaging and predicting treatment response and outcome in oncologyWednesday, March 1, 10:30 - 12:00 Room: M 3 Session Type: Scientific Session Topics: Imaging Methods, Oncologic Imaging Moderators: A. Fohlen (Caen/FR), J. Sosna (Jerusalem/IL) Add session to my schedule In your schedule (remove)
Histogram analysis of ADC from whole-body DW-MRI (WB DW-MRI) to predict very early response to chemotherapy in patients with metastatic colorectal cancer (mCRC): preliminary results
Purpose: We evaluated apparent diffusion coefficient (ADC) histogram analysis parameters from WB DW-MRI as very early predictors of response to chemotherapy in patients with mCRC.
Methods and Materials: 11 patients with histologically confirmed mCRC were scanned with WB DW-MRI (b=0, 50, 150, 500, 900 s/mm2) at 1.5T, before and 10.8±2.7 days after commencing chemotherapy. Response was assessed by the final clinical RECIST 1.1. For analysis, patients were grouped as progressive or non-progressive disease. Mean ADC and histogram analysis parameters (skewness, kurtosis, 25th., 50th. and 75th. percentile) were calculated. Parameters were compared between baseline and follow-up scans using a Wilcoxon signed rank test.
Results: 8 patients did not progress and 3 patients progressed on treatment. Skewness and kurtosis were significantly lower at follow-up (2.57±2.65 vs. 0.93±1.09 and 26.6±36.1 vs. 5.9±6.5 with P=0.003 and P=0.001 respectively) for non-progressors, whilst for progressive disease skewness and kurtosis was higher at follow-up (1.16±0.63 vs. 1.35±1.30 and 6.08±3.44 vs. 11.10±14.04 with P=0.89 and P=0.57 respectively). There was no significant difference between baseline and follow-up in the mean ADC and percentiles for any of the groups.
Conclusion: Skewness and kurtosis are ADC histogram parameters that can potentially separate between progressive and non-progressive disease in patients with mCRC, scanned with WB DW-MRI. This promising ‘one-stop-shop’ response assessment methodology is ready for use in a larger trial.
Comparison of imaging response criteria in metastatic melanoma patients treated with immune checkpoint inhibitors: a single institution analysis
Purpose: Immune-related response criteria (irRC) is increasingly used for assessing tumour response in patients treated with immunotherapy. However, it is unclear how response categorisation by irRC is directly compared with other criteria. We applied Response Evaluation Criteria in Solid Tumours (RECIST 1.1), irRC, Choi and revised-Choi response criteria and compared their associated progression free survival (PFS) and overall survival (OS) in melanoma patients treated with immunotherapy.
Methods and Materials: After institutional review board approval, 144 patients (65% male, mean age 59) treated with Ipilimumab (Group 1, n=100) and Pembrolizumab or Nivolumab (Group 2, n=44) were identified. CT images acquired at baseline, 3- and 6-months after treatment were analysed. A radiologist categorised response according to the above four criteria. Patients demonstrating complete response (CR), partial response (PR) or stable disease (SD) were responders (R); patients with disease progression (PD) were non-responders (NR).
Results: For Group 1, all criteria showed significant differences between R and NR for PFS and OS at 3- and 6-months after therapy (log-rank test p<0.05; hazard ratio between 2.7 and 3.6 for PFS, and between 2.2 and 3.4 for OS, across four criteria). For Group 2, all criteria showed significant differences between R and NR for PFS at 3- and 6-months (p<0.05; hazard ratio between 4.2 and 14.4) but not for OS (p>0.05).
Conclusion: The four imaging response criteria were comparable for patient categorisation of R and NR and for stratification of PFS and OS. IrRC may not provide additional information for the majority of patients on immunotherapy.
3D imaging biomarkers: prediction of survival in patients with non-small cell lung cancer brain metastases treated with stereotactic body radiation therapy
Purpose: To investigate the value of 3D-quantitative tissue enhancement as an early imaging biomarker for patient survival in patents with non-small cell lung cancer (NSCLC) brain metastases treated with stereotactic body radiation therapy (SBRT).
Methods and Materials: 27 patients with NSCLC brain metastasis were treated with SBRT. Baseline contrast-enhanced MRI (ceMRI) was used for image analysis using the qEASL tool (IntelliSpace Portal V.8, Philips Healthcare). A segmentation-based 3D-quantification was performed in each patient to measure the relative tumour enhancement in each lesion. A cut-off value of 65% lesion enhancement was used to stratify the patient cohort in two groups (<65% and >65% volumetric lesion enhancement). Survival was evaluated using Kaplan-Meier analysis and compared using the Mantel-Cox Log-rang test as well as proportional hazard ratios (HR).
Results: Median OS of the entire population was 6.0 months. The stratification of the cohort according to the 65% cut-off for the relative enhancing tumour volume achieved statistical significance p=0.0452 (HR, 2.2 [95% CI, 1.1-6.8]). Patients with >65% enhancing lesion volume survived significantly longer than patients with relatively less-enhancing tumour lesions (8.15 months vs. 2,09 months [HR 0.25; 95% CI, 0.11-0.58], respectively). The total tumour volume did not achieve a statistically significant separation of survival curves.
Conclusion: As opposed to the total tumour which currently is the most commonly used staging marker, volumetric assessment of metastatic NSCLC lesion enhancement on baseline ceMRI is strongly associated with patient survival after SBRT. Specifically, patients with hyper-enhancing lesions demonstrated improved survival as compared to those with hypo-enhancing lesions.
DWI and ADC in assessing early response to angiogenesis inhibitors in metastatic renal cell carcinoma
Purpose: Angiogenesis inhibitors have a potential role in treating metastatic renal-cell carcinoma, but it is still not clear why only some patients can benefit of this treatment protocol. Objective was to look for DWI parameters able to identify early response to angiogenesis inhibitors in patients with metastatic renal-cell carcinoma, considering RECIST1.1 as reference Standard.
Methods and Materials: We prospectively enrolled twenty-four patients candidate to start angiogenesis inhibitors with at least one target lesion and who underwent 1,5T MRI examination with multiple b-values DWI sequences (0,40,200,300,600) in specific timeline: one week before (T0), 2 weeks after (T2) and 8 weeks after (T8) treatment beginning. ADC value was calculated drawing ROIs on the lesions. Twenty-one patients with 30 lesions had adeguate data for comparative evaluation.
Results: At T2 six patients had early response. At T8 4 patients had partial response (PR), 15 disease control (DC), 2 progression disease (PD); average progression free survival was 273 days. PR group, as compared to DC or to PD showed higher T0 ADC values at b40 (respectively 5,5(1,4),3,6(1,2), 1,7(0,9) avg(SD)): we can assess that more vascularised lesions are more responsive to treatment. PD group have significantly lower ADC values then both other groups, at T0, T2 and T8, for all b-values.
Conclusion: Results show that ADC at T0 may help selecting patients with promising good response to angiogenesis inhibitors. Moreover at T0 and at T2 ADC has the potential to select patients who wouldn't benefit from treatment (early progression disease).
CT texture analysis as a predictor of response to therapy and prognosis in patients with metastatic renal cell carcinoma treated with first-line tyrosine kinase inhibitors
Purpose: To investigate if texture analysis (TA) on pretreatment contrast-enhanced CT (CECT) images can predict response to tyrosine kinase inhibitors (TKI) and prognosis in patients with metastatic renal cell carcinoma (mRCC).
Methods and Materials: 65 pretreatment CECT studies of mRCC patients treated with first-line TKI were retrospectively reviewed. Objective response was assessed every 3 months according to RECIST 1.1 and modified Choi (mChoi) criteria. TA was performed on a 5-mm-thick central slice for each target lesion using a commercially available software (TexRAD Ltd, UK). Primary texture features and a novel HeteroDensity Index (HDI), accounting for size-standardised heterogeneity and mean pixel density, were quantified using different spatial-scale filters (ssf). Per-patient texture features were correlated with objective response, progression-free and overall survival (PFS, OS) using logistic regression and survival analysis, statistical significance was corrected to control false discovery rate.
Results: Primary texture features were not able to discriminate responders and non-responders. HDI obtained with a 3-mm ssf (ssf3) was positively correlated with objective response (odds ratio 0.14 for RECIST and 0.2 for mChoi criteria, p 0.018 and 0.026, respectively). Low ssf3 HDI was associated with worse PFS (hazard ratio 4.14, p 0.0001) and OS (hazard ratio 3.36, p 0.0008).
Conclusion: TA on pretreatment CECT helps to predict objective response and prognosis in mRCC patients treated with first-line TKI.
DCE-MRI to assess pathological response to neoadjuvant chemotherapy for high-grade soft tissue sarcomas
Purpose: Aim of this pilot study is to determine if DCE-MRI parameters or their changes show significant differences between good and poor responders to neoadjuvant chemotherapy in high-grade soft tissue sarcomas.
Methods and Materials: 23 patients underwent DCE-MRIs after 2 courses of chemotherapy (MRI-1) and at the end of neoadjuvant protocol, before curative surgery (MRI-2), in our tertiary referral center. 17/23 patients had an available initial MRI (MRI-0). MRIs were performed on two 1.5T-MR-systems with same acquisition parameters and post-processing. Surgical specimens pathological analysis was the gold standard: good response was defined as <10% of viable tumour (11/23) and poor response as ≥50% (12/23). Morphological features on conventional sequences and their changes were reported by two radiologists. For DCE-MRI analysis, ROIs were manually and consensually drawn on tumoural tissue component and on healthy muscle to obtain ratio for each parameter. The wash-in (WI), peak-enhancement (PE), area-under-curve 180s (AUC180) were calculated for semi-quantitative analysis using Olea software. The volume transfer constant (Ktrans) was calculated using extended-Tofts-Kety model. Permeability-curves and AUC180 density-histogram were obtained and scored according to an 8-points qualitative scale.
Results: WI-ratio, PE-ratio, AUC-ratio and Ktrans-ratio on MRI-2 were significantly decreased in good responders (respectively, p=0.04, 0.015, 0.017, 0.018) while differences did not reach significance on early MRI-1. Evaluation of permeability-curve and AUC180 histogram showed early and late significant modifications (p=0.025 and 0.0005).
Conclusion: Qualitative, semi-quantitative and quantitative DCE-MRI may be helpful as an additional technique in evaluating response to chemotherapy in high-grade soft tissue sarcoma.
Response to neoadjuvant chemotherapy of skeletal-osteosarcoma/Ewing sarcoma on the basis of MRI 18-FDG-PET by correlating with pathological necrosis
Purpose: To establish pre/post-chemotherapy MRI and PET-CT scans are as effective as histopathological assessment in assessing response to neoadjuvant-chemotherapy.
Methods and Materials: 34 biopsy proven patients of osteosarcoma/Ewing sarcoma prospectively studied pre/post-chemotherapy for response evaluation by MRI and PET-CT in P. D. Hinduja Hospital-Mumbai. Comparison by assessing post-chemotherapy response in these patients using dynamic contrast MRI/DW-MRI. Post-chemotherapy scans analysed by radiologist and nuclear-medicine specialist without knowledge of neoadjuvant-chemotherapy status.
Results: Paired sample t-test used to analyse change in percentage volume of tumour in pre/post-chemotherapy patients. Mean ADC values in pre/post-chemotherapy were 0.8964 and 2.8043. ADC values in post-chemotherapy responders and non-responders were 3.5437 and 1.8536. Positive correlation between necrosis rate and increased ADC values in post-chemotherapy patients was observed. In post-chemotherapy patients positive correlation was observed between percentage change of volume and percentage change of SUV value. Change in DCE curve types in post-chemotherapy patients were evaluated using chi square test.
Conclusion: Advanced MRI techniques complement standard MRI but remain insufficiently used. DWI/DCE sequences are significantly effective in preoperative response evaluation to chemotherapy in patients with osteosarcoma/Ewing sarcoma patients. Conventional MRI not effective in response evaluation to chemotherapy as much as PET-CT. Preoperative response evaluation by non-invasive imaging modality like MRI helps deciding management and surgical procedure to be either limb salvage or amputation.
Can diffusion-weighted MR imaging be used for prediction and monitoring of treatment response in gall bladder carcinoma?
Purpose: Response evaluation in gall bladder carcinoma (Ca GB) is traditionally done after completion of 3-4 cycles of chemotherapy according to size criteria based on Response Evaluation Criteria in Solid Tumours (RECIST). We hypothesised that apparent diffusion coefficient (ADC) value can be used as an early marker of response to chemotherapy in Ca GB.
Methods and Materials: In this study 23 patients were evaluated. MRI was done at baseline, after 1st (early time point) and 3rd cycle of chemotherapy. Percentage change in size after 3 cycles of chemotherapy was taken as the gold standard to assess response based on RECIST. Total 22 primary masses, 32 lymph nodes and 18 liver metastases were evaluated. Baseline ADC was compared between responders and non-responders. To assess response to chemotherapy percentage ADC change at early time point was compared between responders and non-responders.
Results: Mean percentage increase in ADC at early time point for primary mass, lymph nodes and liver metastases was significantly higher in responders (24.8%, 29.6%, 15.9%) than non-responders (11.5%, 6.9%, 0.8%) respectively (p= 0.02, 0.0001, 0.04). There was no significant difference between baseline ADC of responders and non-responders in primary mass (p= 0.46). Area under curve (AUC) for percentage change in ADC for differentiating responders at early time point was 0.79 for primary mass, 0.89 for lymph nodes 0.79 for liver metastases respectively.
Conclusion: Increase in ADC can be used for early assessment of response to chemotherapy in Ca GB. However baseline ADC cannot predict response to chemotherapy.
Purpose: Imaging in pancreatic cancer is a challenge regarding therapy response evaluation. Tumour size, attenuation, and perfusion are widely used as measurable parameters, but are often limited by blurry tumour borders and missing qualitative parameters. We tested this new CT-based approach of tumour heterogeneity feature analysis for monitoring therapy response.
Methods and Materials: 13 patients with pancreatic adenocarcinoma undergoing portalvenous abdominal computed tomography according to standard as baseline imaging with clinical follow-up and imaging (timespan median 64 days) under systematic therapy (FOLFIRINOX/Gemcitabin) were retrospectively analysed. Progression was defined as new lesions (6 patients) and local tumour spread (3 patients). Tumour heterogeneity analysis was performed using mintLesion®. Image features included tumour volume, entropy, kurtosis, mean positivity of pixels (MPP), skewness, uniformity of distribution of positive pixels (UPP), and uniformity of pixels. Statistical analysis were performed with spearman's rank correlation and Mann-Whitney-U-test.
Results: During follow-up tumour volume did not significantly change between our groups with overall progression (local and systemic) and progression free patients (p=0,661). MPP values were significantly higher in patients without progression compared to patients with progression (p=0.030). There was a significant negative correlation between changes in kurtosis and the time till local tumour spread (p=0.008) or systemic progression (p=0.017).
Conclusion: Results suggest that analysis of tumour heterogeneity may provide valuable information from routine acquired images regarding therapy response evaluation; this may help adjusting therapy regimes and is easily integrated in clinical workflows. Furthermore, this procedure may predict therapy response and, hence may regarded as a marker for progression-free survival.
Sarcopenia is an independant prognostic factor for poor overall survival among patients with pancreatic adenocarcinoma
Purpose: Our purpose was to evaluate sarcopenia as an independent prognostic factor for overall survival in patients with pancreatic adenocarcinoma to predict early relapses or progressive disease and to help decision-making for treatments.
Methods and Materials: From 2009 to 2015, all the patients with pancreatic adenocarcinoma were retrieved (N=114). A retrospective review of the total psoas area (TPA) was performed for each scan (N=713). The TPA was measured on a single cross-sectional image through the third lumbar vertebrae normalised for stature. The percentage decrease in the value of the TPA during the follow-up was also analysed.
Results: In univariate analysis, a TPA level under 420 during the follow-up whatever the time of the measure HR=3.419 ([2.168;5.394]; 95% CI; p<0.0001) and a TPA decrease of more than 20% from baseline HR=7.169 ([4.526;11.353]; 95% CI; p<0.0001) were independent prognostic factors of death. The multivariate analysis, confirmed the results with HR=5.799 ([3.418; 9.839]; 95% CI; p<0.0001) in the non-surgical group and HR=8.089 ([2.157; 30.339]; 95% CI; p=0.0019) in the surgical group for TPA decrease of more than 20% from baseline during the follow-up.
Conclusion: Our analysis shows that regardless of the time of the measurement, a TPA value under 420 and a TPA decrease of more than 20% during the follow-up are strong and independent death risk prognostic factors in patients with pancreatic adenocarcinoma. We now provide a tool to evaluate patients at diagnosis or during iterative consults of treatment or surveillance.