Preoperative local staging of prostate cancer : aspects on predictive models, magnetic resonance imaging and interdisciplinary teamwork

In prostate cancer surgery the two issues at stake are the removal of the tumour on one hand and functional outcome i.e. urinary continence and sexual function on the other. A nerve preserving procedure will optimise the functional outcome but introduces the risk of positive surgical margins by accidentally leaving small tumour remnants behind, thus risking a poor oncological outcome. Preoperative knowledge of tumour aggressiveness, location and whether local growth is confined to the prostate is of outmost importance for an optimal outcome. Currently available tools that provide the surgeon with preoperative information on which to base the treatment decision and surgical technique are far from perfect. The overall aim of this thesis was to explore ways to improve preoperative local staging of prostate cancer, including the development of a prediction tool and the use of magnetic resonance imaging (MRI) in the decision of surgical method.

In Paper 1 we found that of men who underwent surgery with preoperative characteristics implicating very low risk disease, one third had adverse pathology outcome i.e. non‐organ confined tumours and/or more aggressive tumour features at pathology. Sixteen percent had positive surgical margins and only 40% were urinary continent and sexually potent 12 months after surgery. The findings describe both the shortcomings of the preoperative work‐up and the risks linked to surgery. It also gives support to active surveillance, where active treatment is deferred, as an option for men with very low risk, albeit after careful risk stratification where MRI should play an important role to rule out maleficent tumours.

Patients with tumour that on clinical examination are classified as organ‐confined will in approximately one third of the cases subsequently be reclassified at pathology as non‐organ confined. In Paper 2 the development of a prediction tool from preoperative variables, predicting non‐organ confined disease, is described. The accuracy of the final model was only moderate and when validated on an external group showed even lower performance. We found that the probable cause of the low performance was due to variability between pathologists in judgement of our primary outcome measure, tumour stage. This underlines the need for validation before the use of an externally derived prediction model.

Paper 3 investigated the additional value of a three‐dimensional (3D) T2‐weighted sequence with radial reconstructions, in local staging of patients receiving a preoperative prostate MRI. A radial reconstruction overcomes the partial volume effect encountered at the curved portions of the prostate seen with conventional imaging methods. The outcome however showed no benefits of adding the 3D sequence but rather introduced an uncertainty when comparing assessments from two radiologists, with an inter‐rater correlation of 0.17 (poor agreement) compared to traditional sequences of 0.42 (moderate agreement).

In Paper 4 we compared outcome measures from pathology regarding positive surgical margins between (A) men who had performed a preoperative MRI discussed at an interdisciplinary consensus conference between surgeons and a radiologist and (B) men who were operated on without a preoperative MRI. The group receiving MRI and a conference showed a significant reduction in positive surgical margins but at the cost of less nerve sparing procedures, compared to those men not receiving a preoperative MRI.

The findings of this thesis highlight the difficulties encountered at prediction of local tumour stage in prostate cancer at all stages of the preoperative investigation. This implicates the need for improvements, with tuning and standardisation of the different preoperative investigational modalities for better oncological and functional outcome in men undergoing treatment with curative intent. This should be carried out in a multi‐disciplinary setting, to optimize and increase the knowledge of all specialists involved in the care of prostate cancer patients.

List of scientific papers

I. Oncological and functional outcomes one year after radical prostatectomy for very low risk prostate cancer. Results from the prospective LAPPRO trial. Carlsson S, Jäderling F, Wallerstedt A, Nyberg T, Stranne J, Thorsteinsdottir T, Carlsson SV, Bjartell A, Hugosson J, Haglind E, Steineck G. British Journal of Urology Int. 2016 Aug;118(2):205‐12.
https://doi.org/10.1111/bju.13444

II. Accurate prediction tools in prostate cancer require consistent assessment of included variables. Jäderling F, Nyberg T, Blomqvist L, Bjartell A, Steineck G, Carlsson S. Scandinavian Journal of Urology. 2016 Aug;50(4):260‐6.
https://doi.org/10.3109/21681805.2016.1145736

III. Accuracy in local staging of prostate cancer by adding a three‐dimensional T2 weighted sequence including radial reconstructions in magnetic resonance imaging. Jäderling F, Nyberg T, Öberg M, Carlsson S, Skorpil M, Blomqvist L. [Submitted]

IV. Preoperative staging using magnetic resonance imaging and risk of positive surgical margins after prostate cancer surgery. Jäderling F, Akre O, Aly M, Björklund J, Olsson M, Adding C, Öberg M, Blomqvist L, Nyberg T, Wiklund P, Carlsson S. [Submitted]