MRI scoring system predicts post-surgery prostate cancer recurrence
MedWire News: A novel scoring system for magnetic resonance imaging (MRI) findings strongly predicts biochemical recurrence after radical prostatectomy, potentially improving recurrence risk assessment, conclude US scientists.
The selection of treatment options for patients diagnosed with prostate cancer depends on accurate assessment of the stage and extent of disease and the likelihood of recurrence. As the current nomograms based on clinical variables have substantial limitations, it is important that additional tools are developed.
To examine the utility of converting MRI findings into a scoring system, Hedvig Hricak, from Memorial Sloan-Kettering Cancer Center in New York, and colleagues studied 610 patients with biopsy-confirmed prostate cancer who underwent MRI before radical prostatectomy and whole-mount step-sectioning of the pathology sample.
The MRI findings were scored using a seven-point system based on the TNM staging system: 1, no tumor seen; 2, tumor seen, no extracapsular extension (ECE); 3, tumor seen, ECE cannot be excluded; 4, unilateral ECE; 5, bilateral ECE; 6, seminal vesicle invasion; and 7, lymph node metastasis. The results were then combined with previously published pre-operative nomogram variables.
In all, 10.5% of patients experienced biochemical recurrence. The median follow-up period was 37.9 months, at 40.2 months for those with no recurrence, and 18.5 months for those with recurrence. The median pre-operative serum prostate-specific antigen (PSA) level for patients with and without recurrence was 12.65 ng/ml and 6.63 ng/ml, respectively.
The team notes in the BJU International that 0.3% had an MRI score of 1, while 69.8% scored 2, 11% scored 3, 14.3% scored 4, 1.5% scored 5, 2.1% scored 6, and 1% scored 7. Actuarial recurrence rates for each score were 0%, 4.5%, 9%, 24.1%, 33.3%, 69.2%, and 100%, respectively.
Taking into account PSA, clinical stage, biopsy Gleason score, and biopsy core data, MRI scores were significantly associated with 5-year and 10-year biochemical recurrence, at hazard ratios of 1.76 and 1.81, respectively.
Adding MRI scores to the nomogram increased the concordance indices of the 5- and 10-year models from 0.762 to 0.776 and from 0.773 and 0.788, respectively, neither of which increase was significant.
“Although MRI did not provide added prognostic value to standard clinical nomograms, in centers where pretreatment MRI is used routinely, it might increase the confidence of the clinician in assessing risk by contributing supporting data,” the researchers write.
MedWire (www.medwire-news.md) is an independent clinical news service provided by Current Medicine Group, a part of Springer Science+Business Media. © Current Medicine Group Ltd; 2009
By Liam Davenport