Many of these patients identified with small volume recurrence in nodal tissue are now being offered salvage lymphadenectomy as a treatment option – at the very least, it may delay time to hormone therapy. However, as with all new technologies, with increased utilization users have noted potential discrepancies between PET findings and clinical findings.
In this series, the authors of multiple institutions with significant experience with salvage lymphadenectomy based on PSMA PET imaging retrospectively compare PET imaging findings vs. final surgical pathology in 525 men who received extended salvage lymphadenectomy for oligometastatic disease. All patients were diagnosed with PET/CT scan using either 11C-choline (n=356; 68%) or 68Ga-PSMA (n=169; 32%). They hypothesized that PET imaging underestimated the volume of metastatic disease. As such, their primary outcome was the number of underestimated nodes: number of positive lymph nodes at SLND minus number of positive spots at PET/CT scan. The variables they assessed for association with underestimation were: the number of positive spots on PET/CT scan, PSA level at the time of SLND, PET/CT tracer (11C-choline vs. 68Ga-PSMA), and the number of nodes removed.
Of the 525 men, the number of positive spots on pre-operative PET/CT scan was 1 in 364 men (52%), 2 in 151 men (22%), 3 in 101 men (15%), and ≥4 in 79 men (11%) patients, respectively. Interestingly, most studies would consider those with 4+ nodes outside the oligometastatic space, so the fact that they got an SLND is indicative of the aggressive nature of these groups. This is highlighted by the fact that ~60% had pT3 disease, 38% were Gleason 8-10, 34% had positive surgical margins, and 20% were pN1 at the time of RP. 28% of patients had detectable PSA post-RP.
The median number of lymph nodes removed at SLND was 19 (IQR: 10, 31). The time from RP to SLND was 28 months on average, so approximately 2 years.
The number of positive nodes was 0 in 110 men (16%), 1 in 165 men (24%), 2 in 82 men (12%), and ≥3 in 338 (49%) patients, respectively. Median under-estimation of tumour burden was 1 (IQR: -1, 4).
At multivariable analysis, accounting for the other variables mentioned (modality of PET scan, PSA level at time of SLND, and number of nodes removed at SLND), the number of positive spots at PET/CT scan was a significant predictor of under-estimation of tumour burden (linear coefficient: 0.23; p=0.02).
The interaction test for the hypothesis that under-estimation of tumour burden varies by PET/CT tracer was statistically significant (p<0.0001). Under-estimation was relevant with both tracers, but it was more evident with 11C-Choline as compared to 68Ga-PSMA.
Ultimately, the more spots present on the PET scan, the higher the chance of finding higher volume disease at the time of SLND. While the authors conclude that this supports extended SLND, I would argue that it would imply that higher volume disease patients may not benefit from SLND as they may have disease elsewhere that is not being removed by SLND – which is why many of these patients with 4+ nodes maybe should not have been considered oligometastatic!
This obviously warrants further discussion, but it is reassuring to know that PET scans with 1-2 spots are less likely to be underestimating the volume of disease.
Some questions that came up during the discussion:
1. Was there are a difference in patients who did not have a node dissection initially (pNx) vs. those who were pN1? The authors state there was no significant difference.
2. The authors made it clear during the discussion that this should be taken in the context of clinical data, not in isolation – PET scan still has value, just not in isolation!
Presented by: Nicola Fossati, MD
Co-Authors: Nazareno Suardi, Giorgio Gandaglia, Carlo Andrea Bravi, Milan, Italy, Matteo Soligo, Jeffrey R. Karnes, Rochester, MN, Maximilian Schmautz, Axel Heidenreich, Cologne, Germany, Annika Herlemann, Christian Gratzke, Christian Stief, Munich, Germany, Antonino Battaglia, Wouter Everaerts, Steven Joniau, Hendrik Van Poppel, Leuven, Belgium, Almut Kalz, Daniar Osmonov, Klaus-Peter Juenemann, Campus Kiel, Germany, Nieroshan Rajarubendra, Inderbir S. Gill, Los Angeles, CA, Shahrokh F. Shariat, Vienna, Austria, Alexander Mottrie, Melle, Belgium, Francesco Montorsi, Alberto Briganti, Milan, Italy
Written by: Thenappan Chandrasekar, MD, Clinical Fellow, University of Toronto, | twitter: @tchandra_uromd at the 2018 AUA Annual Meeting - May 18 - 21, 2018 – San Francisco, CA USA