Evaluation of Diffusion-Weighted MRI and FDG-PET/CT to Assess Response to AdCD40L treatment in Metastatic Melanoma Patients

Abstract

The purpose was to evaluate the potential of diffusion-weighted-magnetic resonance imaging (DW-MRI) and ¹⁸F-fludeoxy-glucose-positron emission tomography integrated with CT (FDG-PET/CT) for prediction of overall survival (OS) following AdCD40L-immunotherapy in patients with metastatic malignant melanoma (MMM). Twenty-four patients with refractory MMM were treated with immunostimulatory AdCD40L gene therapy in a phase I/IIa study. Pre-therapeutic DW-MRI and FDG-PET/CT were performed and then repeated at 5 and 9 weeks post-treatment. Evaluation was conducted according to RECIST 1.1 and EORTC criteria. Apparent diffusion coefficient (ADC), true diffusion coefficient (D), maximum standardized uptake value (SUV_max) were measured in the injected lesions. Fold changes (F) in ADC (F ADC), D (F D), SUV_max (F SUV_max) were statistically assessed. F D ≥ 1 and F ADC ≥ 1 were associated with better OS in scans at week 5 and 9 respectively. F SUVmax was not correlated to OS. F ADC ≥ 1 in both post-treatment scans and F D ≥ 1 at week 5 were related to a significant decrease of size of the injected lesions. These results suggest that in patients with MMM treated with AdCD40l, functional parameters of DW-MRI are better early predictors of OS than the established metabolic and morphologic criteria for FDG-PET/CT and MRI, respectively.

Figure 1

Overall survival (OS) of patients with metastatic melanoma treated with AdCD40L gene therapy assessed by fold change (F) F ADC, F D and F SUV_max of the injected metastases. The analyses were conducted with inclusion of all patients who underwent at least one post-treatment examination (FDG-PET/CT or DW-MRI, n = 21). The Kaplan-Meier method was used followed by log-rank tests. (a) The median OS of the patients with F D ≥ 1 (black line) compared to the median OS of the patients with F D < 1 (grey line) in Scan 1 at week 5 (p = 0.013). (b) The median OS of the patients with F D ≥ 1 (black line) compared to the median OS of the patients with F D < 1 (grey line) in Scan 2 at week 9 (p = 0.032).(c,d) The median OS between patients with F ADC ≥ 1 (black lines) and compared with those with F ADC < 1 (grey lines) in Scan 1 at week 5 (c; p = 0.251) and in Scan 2 at week 9 (d; p = 0.057).

Figure 2

Overall survival (OS) of patients with metastatic melanoma treated with AdCD40L gene therapy assessed by fold change (F) F ADC, F D and F SUV_max of the injected metastases. The analyses were conducted excluding the patients who did not undergo all examinations (n = 13). The Kaplan-Meier method was used followed by log-rank tests. (a) The median OS of patients with F D ≥ 1 (black lines) compared with those with F D < 1 (grey lines) observed in Scan 1 at week 5 (p = 0.075).(b) The median OS of the patients with F D ≥ 1 (black line) compared with the median OS of the patients with F D < 1 (grey line) in Scan 2 at week 9 (p = 0.001).(c) Comparison of the median OS in patients with F ADC ≥ 1 (black lines) and those with F ADC < 1 (grey lines) in Scan 1 at week 5 (p = 0.15).(d) Comparison of the median OS of the patients with F ADC ≥ 1 (black line) to the median OS of the patients with F ADC < 1 (grey line) in Scan 2 at week 9 (p = 0.014). (* indicates p < 0.05).

Figure 3

Correlation analysis of overall survival (OS) with D, and subsequently ADC, in the injected metastases. The Pearson’s correlation analysis was used. (a) F D correlation with OS at week 9 (Scan 2) including all patients (n = 21, p = 0.015, R² = 0.378). (b) F ADC correlation with OS at week 9 (Scan 2), including all patients (n = 21, p < 0.01, R² = 0.628). (c) Correlating OS in the group of patients that were “responders” in our previous publications^, based on PET/CT evaluations (n = 6) with fold change (F) F ADC in Scan 2 at week 9 (p = 0.023, R² = 0.861). (d) F f values in the injected metastases were correlated to F SUV_max values at week 9 (Scan 2), including all evaluable patients (n = 21, p = 0.01, R² = 0.442). The Pearson’s correlation analysis was used.

Figure 4

(a) Patients with F D value above the median had a significantly better OS at week 5 (Scan 1), including all patients (n = 21, p = 0.011). The black line shows the F D value above the median, while the grey line shows F D value below the median.(b) Correlation for F ADC with the fold change of the lesion size based on RECIST 1.1 at week 5 (Scan 1). The gray line is the fitted regression line. The correlation coefficient was −0.62 and p = 0.008. (c) Correlation for F ADC with the fold change of the lesion size based on RECIST 1.1 at week 9 (Scan 2). The gray line is the fitted regression line. The correlation coefficient was −0.68 and p = 0.016. (d) Correlation for F D with the fold change of the lesion size based on RECIST 1.1 at week 5 (Scan 1). The gray line is the fitted regression line. The correlation coefficient was −0.49 and p = 0.044. (* indicates p < 0.05).

Figure 5

DW- and PET-images of a 67-year-old woman with disseminated mucosal melanoma in vulva (cohort 2). Four intratumoral injections with AdCD40L were given to a right inguinal lymph node metastasis. The patient had partial metabolic response (PMR) in the first and stable metabolic disease (SMD) in the second post-therapy evaluation, according to EORTC criteria while the case was assessed as stable disease (SD) at both post-therapy scans according to RECIST 1.1. An increase in ADC/D in the injected metastasis was observed at both DW-MRI scans post-therapy. The short axis of the injected metastasis was measured at each time date and was unaltered according to the RECIST 1.1 criteria. An initial decrease of the f% value was observed while it was increased at the second DW-MRI scan post-therapy. In a not injected left inguinal lymph node metastasis the ADC-value was increased at both DW-MRI scans post therapy while the value of D was initially increased but it was decreased in the second post-therapy evaluation. A similar pattern as for the D-value was observed regarding the values of f% while the size of the metastasis was unaltered. A third metastasis at the proximity of the urinary bladder is present in the DW-images. This metastasis was not distinguishable at the PET/CT scans due to the high activity of the urinary bladder. (a–c) Diffusion-weighted MR image (DWI) at b = 900 s/mm² -in the axial plane. The arrows indicate the injected right inguinal metastasis at baseline (a), in Scan 1 at week 5 (b) and in Scan 2 at week 9 (c). (d–f) FDG-PET images. The arrows indicate the injected right inguinal metastasis at baseline (a), in Scan 1 at week 5 (b) and in Scan 2 at week 9 (c).