Limitations in assessing response in CRPC have long been recognized. Neither PSA nor derived measures such as PSA change and PSA velocity have met surrogacy criteria for survival benefit [50]. Despite this, PSA remains a common measure of treatment response [41]. PSA is widely used in Phase I/II trials, although the mechanism of action of the therapy may render PSA an unhelpful marker of activity. For example, immunotherapies may demonstrate improved survival without impacting PSA kinetics [52, 53].

In the 1970s bone scintigraphy was adopted for the detection of bone metastases [54]. False negative scans due to lack of osteoblastic response or small lesion size, lack of specificity, and difficulties with reproducibility were quickly recognized [55]. Despite these issues, bone scans became part of the standard staging for men with prostate cancer. The PCWG2 recommended assessing progression based on the appearance of new lesions, rather than more subjective measures relating to tracer intensity, or individual lesion size or area [41]. More recent attempts to improve bone scan assessment have included computer-aided interpretation, which showed a good ability to reproduce the bone lesion area, intensity, and count assessments of a highly experience nuclear medicine physician [56].

The Response Evaluation Criteria in Solid Tumours (RECIST) were published in 2000 and aimed to standardize response assessment for soft tissue disease manifestations [57]. These criteria were incorporated in a modified form in the PCWG2 recommendations, including 3-monthly restaging with CT and bone scans. Alternative imaging modalities have been proposed and although the optimal modality and technique remains unclear, there is increasing enthusiasm for replacing the traditional methods of prostate cancer assessment. Whole-body DW-MRI has been shown to outperform Tc 99 m bone scanning and may be suitable to assess treatment responses [58]. 18 F-Choline positron emission tomography-CT (PET-CT) may also offer improved sensitivity and specificity in detection of prostate cancer metastases [59]. However more data are required to bring these imaging modalities into routine clinical practice and trial design.

Novel biomarkers for efficacy monitoring include CTC counts and biology-driven markers of treatment effect [50]. These strategies require a rigorous process to prove surrogacy, but can be useful to identify patients likely to benefit from targeted therapy. Encouraging results were observed in the COU-AA-301 trial, where CTCs in combination with LDH demonstrated survival surrogacy on an individual patient level [60, 61].