Synopsis

Lesion size threshold is the most common imaging feature used to assess response to therapy. Size as an imaging feature has its limitations. Quantitative imaging biomarkers (QIBs) could identify subtle microstructural changes prior to morphological changes. In this study, we explored the use of novel whole-body MRI (WB-MRI) QIBs for nodal disease characterisation and treatment response monitoring in radio-recurrent prostate cancer (rPC). We showed signal fat fraction could discriminate between positive and negative nodes and that it can be used for response monitoring.

Introduction

Whole-body MRI (WB-MRI) offers the opportunity to provide a cost effective solution for cancer staging (1). However, attempts to use quantitative imaging biomarkers (QIBs), such as apparent-diffusion-coefficient ADC, for classification and response monitoring have met with limited success [2,3] and are rarely applied in the clinical arena. Lymph nodes are composed of a predominantly fatty hilum surrounded by a cellular rim [4]. Replacement of the fatty components of lymph node by cancer cells is commonly seen in patients with metastatic prostate cancer [5]. Recent studies in multiple myeloma have shown that mDixon signal fat-fraction (sFF) is a repeatable and useful QIB for classification and treatment response (6). In this study, we explored the value of sFF as a nodal status classifier and response marker in patients with radio-recurrent prostate cancer (rPCa).

Methods

Patients with suspicion of rPCa as per phoenix criteria (7) underwent prostate multi-parametric MRI, 18F-choline-PET-CT, 99mTc bone scan as routine imaging. For research purposes, a baseline 3T WB-MRI was performed (fig 1). Patients underwent treatment guided by routine imaging. All patients were re-imaged with follow-up WB-MRI at 1 year. Of 120 patients with 1-year follow-up WB-MRI, 40 were randomly selected for quantitative analysis. Up to10 anatomically distributed nodes were selected for analysis for each patient. sFF for each node was derived as previously described (8) : SIpre-contrast-F / (SIpre-contrast-F + SIpre-contrast-w ). An enhanced-reference-standard (ERS) fig.2 was applied to each node (using a combination of choline PET-CT, nodal size change between baseline and follow-upWB-MRI (See Fig 3) and PSA kinetics .Nodes were classified as positive, negative or unknown based on the ERS. Baseline WB-MRI node size (short axis diameter)and sFF were compared between positive and negative using the Mann-Whitney test. Receiver-operating-characteristic (ROC)area-under-curve (AUC) calculated for assessment of sFF as a classifier of nodal disease status. Positive nodes were identified in 14/40 patients(fig 4). 13/14 patients were treated with androgen deprivation therapy (ADT). Patients were divided into responder and non-responder groups based on PSA PCWG-2 response criteria (9). Baseline and follow-up (post-treatment) sFF of nodes within these 13 patients was extracted. Baseline size and sFF, and percentage change in (delta)sFF after treatment were compared between responder and non-responder groups using the Mann-Whitney test. ROC analysis of (delta)sFF was performed.

Results

40 patients [median age 73 (range 65-85years) and median PSA 4.45 (range 0.24-28.3ug/L)] were identified. A total of 206 nodes across the 40 patients were analysed(per patient median number 4, range 1 to10). 75/206 were classified as unknown and excluded from sFF analysis. 30/131 were positive (short axis median size 0.8 cm, range 0.6-2.7 cm) and 101 (short axis median size 0.7 cm, range 0.3-1.2 cm) negative by ERS. As expected, positive nodes were significantly larger than negative nodes (p=0.03). The median sFF was significantly lower for positive (0.63) compared withnegative nodes (0.79) (p<0.0001)(figure 5a).Nodal sFF ROC-AUC was 0.86 for classification of metastatic nodal disease. 28 nodes were included from 13patients for treatment response analysis(fig4).10/13 patients responded to treatment and 3/13 were non-responders by PCWG-2PSA criteria; providing 21nodes in the responder and 7nodes in the non-responder groups respectively. sFF nodal characteristics for the responder and non-responder nodes are given in fig5c. Baseline median sFF was lower for nodes in the responder(0.58a.u.) compared with non-responder(0.75a.u.) group(p<0.0001). Median (delta)sFF (following ADT) was 31% (range 6.1to 230.9) in the responder and -6.5% (range -24.2 to 20.6) in the non-responder group(p<0.01), see fig 5d. The ROC-AUC of (delta)sFF was 0.85.

Discussion

Our results highlight the potential of sFF as a marker of nodal disease status and treatment response in patients with rPCa. Prior work demonstrated that sFF measurements is highly repeatable and can be used to assess bone lesions (10). Here we show that sFF can provide a quantitative imaging biomarker able to(ROC-AUC of 0.85) to classify pre-treatment nodal disease status. Furthermore, we show that (delta)sFF may help identify non-responding lymph nodes (ROC-AUC 0.86). We derived an ERS, employing follow-up WB-MRI and PSA and accounting for limitations in performance of routine imaging tests. Our initial results appear promising, prompting on-going analysis of the remaining 80 patients with follow-up WB-MRI. Future work will focus on assessing the value of sFF in bone lesions and finally in clinical decision making for patients with rPCa.

Conclusion

sFF has value as a quantitative imaging biomarker with potential to classify lymph nodes disease status and monitor treatment response in patients with rPCa

Acknowledgements

References

References

1. Padhani ARet al. Rationale for Modernising Imaging in Advanced Prostate Cancer. Eur Urol Focus. 2017.

2. Seber Tet al. Diagnostic value of diffusion-weighted magnetic resonance imaging: Differentiation of benign and malignant lymph nodes in different regions of the body. Clin Imaging. 2015. doi:10.1016/j.clinimag.2015.05.006

3. Roy Cet al. Value of diffusion-weighted imaging to detect small malignant pelvic lymph nodes at 3 T. Eur Radiol. 2010

4. Skandadas G et al. Nodal staging. Cancer imaging. 2009;9(1):104-111

5. Thoeny HC et al. Metastases in Normal-sized Pelvic Lymph Nodes: Detection with Diffusion-weighted MR Imaging. Radiology. 2014.

6. Latifoltojar A et al. Whole body magnetic resonance imaging in newly diagnosed multiple myeloma: early changes in lesional signal fat fraction predict disease response. Br J Haematol. 2017

7. Abramowitz MC et al. The phoenix definition of biochemical failure predicts for overall survival in patients with prostate cancer. Cancer. 2008;112(1):55-60.

8. Messiou Cet al. Assessing response of myeloma bone disease with diffusion-weighted MRI. Br J Radiol. 2012.

9. Scher HIet al. Design and end points of clinical trials for patients with progressive prostate cancer and castrate levels of testosterone: Recommendations of the Prostate Cancer Clinical Trials Working Group. J Clin Oncol. 2008. doi:10.1200/JCO.2007.12.4487

10 Latifoltojar Aet al. Whole-body MRI quantitative biomarkers are associated significantly with treatment response in patients with newly diagnosed symptomatic multiple myeloma following bortezomib induction. Eur Radiol. 2017.