ORLANDO–The value of a molecular marker in cancer depends on whether its use will improve clinical outcomes. Studies on tumor biomarkers are in preliminary stages, and there is a “near total” absence of any papers examining the clinical implications of using markers, according to a speaker at the 2009 Genitourinary Cancers Symposium.
Andrew J. Vickers, PhD, associate attending research methodologist at Memorial Sloan-Kettering Cancer Center in New York, outlined the crucial questions that must be addressed in order to bring good prognostic markers into the clinic.
• Can the marker be measured accurately and reproducibly? “Can you run a blood sample through a mass spectroscopy machine on two different days and get the same spectra?” he said. “Or between two different machines on the same day and get the same spectra? This is an absolutely critical question.”
• Does the marker distinguish between convenient samples of different groups?
“Clearly, if the marker failed to distinguish between obviously different groups, it would not be worth conducting further research,” Dr. Vickers said.
• Is the marker associated with outcomes in the sort of patients to which the marker would be applied in practice?
• Does the marker provide additional information to that already available to the clinician?
“A man’s risk of recurrence after prostate cancer depends on stage, grade, and PSA. A new marker will be useful only if it provides information over and above these variables,” he elaborated.
• Does use of the marker improve clinical outcome?
Answers to the first four questions are rarely sufficient to show that a marker is of clinical value, Dr. Vickers said. Nevertheless, some markers are being aggressively promoted after showing differences between convenience samples or statistical associations with outcomes.
Dr. Vickers cited early prostate cancer antigen (EPCA) and PSA velocity as examples. The EPCA biomarker has yet to be evaluated in men at elevated risk for prostate cancer undergoing biopsy but is still described as a “highly specific serum marker,” Dr. Vickers said. “I have not been able to find a single paper that has been published on EPCA that actually looks at men you would want to give the EPCA test to,” he added.
Similarly, the widespread use of pretreatment PSA velocity is not backed by a single research paper that examined whether using PSA velocity would actually improve outcomes in patients with prostate cancer.
“Pretreatment PSA velocity is pretty much ubiquitous. It’s in the National Comprehensive Cancer Network (NCCN) guidelines. They say, ‘Even if you have a very low PSA, less than 2.5, you should consider biopsy if you have a high PSA velocity’,” he said.
PSA velocity is also used for prognostication, Dr. Vickers continued. “A man treated at Memorial Sloan-Kettering Cancer Center had a high PSA velocity and was told he would need neoadjuvant hormones and should consider being enrolled in a randomized trial of neoadjuvant chemotherapy,” he explained. “He actually had a Gleason 6 tumor and low-volume, organ-confined disease on radical prostatectomy. If you are going to recommend that patients get extremely toxic and dangerous treatments on the basis of a PSA velocity, you have to be quite sure that there is good evidence,” he emphasized.
Dr. Vickers said he read 87 research papers on PSA velocity, but “the number of papers that actually examined whether use of PSA velocity would improve outcome was absolutely zero.”
Ultimately, clinicians need to know if any biomarker will positively affect clinical outcome. “Does getting information on that biomarker change a clinical decision, and does that changed clinical decision improve the outcome for the patient?,” he said. The GU symposium was jointly sponsored by ASCO, the American Society for Radiation Oncology (ASTRO), and the Society of Urologic Oncology.