Read the Full Video Transcript

Rashid Sayyid: Hello, everyone. This is Rashid Sayyid, I'm a urologic oncology fellow at the University of Toronto, and along with Zach Klaaseen, assistant professor and program director at Augusta University. We'll be discussing the recent publication looking at Radical Cystectomy versus Trimodality Therapy for Muscle-Invasive Bladder Cancer: A Multi-Institutional Propensity Score Matched and Weighted Analysis. This article was recently published in Lancet Oncology with Dr. Alexandre Zlotta as the first author.

So we know that radical cystectomy with neoadjuvant chemo in cisplatin-eligible patients is a gold standard approach for the management of muscle-invasive bladder cancer patients. But we also know that radical cystectomy is associated with significant perioperative morbidity and mortality. This is particularly of concern in the muscle-invasive bladder cancer disease state, where bladder cancer patients are often over 65 years of age, have significant comorbidities that come along with their long-term smoking history, and as such, becomes clear that we need organ sparing approaches to maintain the patient quality of life without compromising oncologic outcomes and avoiding the morbidity that comes with the radical cystectomies.

So trimodality therapy or TMT, entails essentially maximal TURBT followed by concurrent chemo radiotherapy and has emerged as a guideline recommended treatment approach for select muscle-invasive bladder cancer patients. Now, we see here the recommendations from the EAU guidelines, the most recent one, looking at the bladder cancer. They do recommend with a strong rating that you may offer surgical intervention or TMT to appropriate candidates as primary curative therapeutic approaches since they are more effective than radiotherapy alone. They also mentioned that you may offer TMT as an alternative to selected well-informed compliant patients, especially for whom radical cystectomy is not an option or not acceptable.

But despite the growing body of evidence regarding safety and efficacy and so endorsement by guidelines, TMT is still not widely used and essentially restricted to poor surgical candidates for whom radical cystectomy is not an option. So level one comparative evidence for TMT versus radical cystectomy is lacking, and ideally this is what we would have in the disease space. However, it's become clear over the last decade that RCTs in this disease space are quite challenging and many have prematurely closed due to insufficient accrual. Now, we see this in the SPARE trial that looked at radical cystectomy versus selective bladder preservation in the UK. This trial was only able to accrue 45 patients in 30 months with still poor adherent to the site treatment strategy.

The BRAVO trial, which was a little bit different, looking at radical cystectomy versus BCG for high risk, high grade, non-muscle invasive bladder cancer, also closed prematurely due to issues with poor accrual. As such becomes quite clear that given that future RCTs in this disease space are unlikely, we need studies such as this that evaluates oncologic outcomes in the large contemporary multi-center cohort of select MIBC patients who would have been eligible for both procedures at high volume centers.

So to this end, the authors conducted a retrospective analysis of patient with clinical T2-T4 node0 M0 bladder cancer patients across three institutions at Toronto, Mass General Hospital, and USC between January, 2005 and December, 2017, and this is the key part here, the eligibility criteria. So this wasn't all-comer MIBC patients but select MIBC patients. This included patients with tumors less than seven centimeters that were solitary. Patients had either no or only unilateral hydronephrosis.

Patients had to have adequate bladder function given that they were going to receive radiotherapy. No extensive or multifocal CIS. If patients had limited CIS adjacent to the primary tumor, that was okay. Histologic variants were okay as long as it was primary urothelial cancer. Patients had to have no contraindications to radiotherapy; for example, prior radiotherapy or inflammatory bowel disease and no concomitant upper tract urothelial cancer.

The decision for TMT versus radical cystectomy in this cohort was based on patient preference following a multidisciplinary discussion of the various treatment options. I want to point out that TMT was only performed at the University of Toronto and Mass General, whereas radical cystectomy was only performed at Toronto USC. So not all two modalities were performed across all three institutions, and there was some variability in what constitute the TMT protocol, especially in the MGH cohort because there were different chemo radiotherapy protocols across different trials at that site.

But I'll give you an example of the TMT protocol Toronto, where all patients underwent a maximal TURBT with repeat resection, performed for cases where tumor was still visible macroscopically, and then patients received concurrent chemo radiotherapy. So the radiotherapy component entailed daily image-guided intensity, modulated radiotherapy, to the bladder, and the pelvic nodes, 46 Gys and 23 fractions, with an additional sequential tumor boost of 20 Gys and 10 fractions for a total of 66 Gys. For the chemotherapy, patients received single agent cisplatin weekly during the therapy.

This data is based on the previous publication JC on 2015. It's important to note that neoadjuvant chemo was offered in Toronto, whereas adjuvant chemo for TMT patient was offered MGH. In Toronto, conversely, for the radical cystectomy patient, they were offered neoadjuvant chemo and adjuvant chemo only if they had pathologic T3, T4 disease, or node-positive disease. For surveillance, all patients underwent cross-sectional imaging as well as cystoscopies to monitor for any intravesical recurrences.

The primary outcome was metastasis-free survival defined as distant or regional pelvic nodal disease. For secondary outcomes, they're looking at overall survival, cancer-specific survival, distant metastatic failure-free survival, which is similar to the metastasis-free survival, so the primary one, but only includes the distant components opposed to the regional as well, and disease-free survival, and recurrences were confirmed either by biopsy or by cross-sectional imaging.

What was the definition of local recurrence? Essentially, it's any urethra or bladder recurrence. The bladder bed in the radical mastectomy group. Ureter recurrences of note were considered a second primary. For the statistical analysis part, the authors use propensity scores to match the radical cystectomy TMT patients. Essentially, the propensity score does. It tries to create or mimic an RCT and create two cohorts of patients in each treatment arm that based on predictive factors, would be equally as likely to receive either treatment.

So essentially what we're trying to do is mimic an RCT model and say that, "Oh, patients would have been randomized to one or the other as likely based on their baseline characteristics." So the way they do this is they calculate the propensity score using logistic regression modeling with a following predictors, age, sex, presence or absence CIS, the clinical T-stage, the performance status, the BMI presence or absence of hydronephrosis, again, unilateral versus not, received peri-treatment chemotherapy and smoking history. The thought that if patients are balanced for these variables, then they would be equally likely to receive one or the other, thus minimizing any potential selection biases.

Propensity scores are incorporated in two ways, a three-to-one matching using the nearest neighbor matching or the inverse probability treatment waiting. After performing the propensity score analysis, the authors further adjusted for these variables by performing a multi-variable Cox proportional hazards model to compare the survival outcomes between TMT and radical cystectomy patients, essentially adjusted for the same variables that they already used to calculate propensity scores.

Of note, they also use competing risks analysis whereby for cancer-specific survival, metastasis-free survival, distant-free survival, distant metastatic failure-free survival, and regional failure-free survival, they use the multi-variable Fine and Gray regression models that accounts for the competing risk of death due to other causes. This generates sub-distribution hazard ratios as opposed to the more traditional hazard ratios that we get with Cox regression modeling. The authors also perform sensitivity analysis to a sense for potential differences in the modality outcomes between the centers, and they also for the propensity score analysis. They also try different ratios for matching, as opposed to three-to-one, they use one-to-one, one-to-two, and one-to-four as well.

Of note, in the TMT cohort, they use time to non-muscle invasive and muscle invasive failures. These were estimated using cumulative incidence functions, which again account for the competing risk of death from other causes. They also looked at the cancer-specific survival stratified by the salvage radical cystectomy status, and they also looked at metastasis-free survival by peri-treatment chemotherapy status. At this point, I'll turn it over to Zach to discuss the results and the discussion conclusion section of this paper.

Zach Klaassen: Thanks so much, Rashid. It's a great introduction for this paper. This is the baseline characteristics before and after matching in this study. So we can see here on the left is before matching, on the right is after three-to-one matching, and just for way of discussion will mostly focus on the after the three-to-one matching for these two cohorts between radical cystectomy and trimodality therapy.

So when we look at age as expected, 71.4 median years of age for these patients. Majority of these patients were male at about three-quarter percentage of these patients. Carcinoma in situ was not present in roughly 77% to 79% of patients. As you'd expect, the majority 90% were clinical T2 patients. Majority of patients were BMI less than 30, and the majority of patients had no hydronephrosis, roughly 88% to 90%. We look at neoadjuvant or adjuvant chemotherapy, I see that just over half of patients had chemotherapy, 59%, 56%. Majority of patients, three-quarters roughly were current or former smokers, and roughly three-quarters of patients had an excellent ECOG performance status of zero.

The next several slides will look very similar to this one. This is the IPTW analysis on the left and the propensity score matching on the right. You can see radical cystectomy in red and trimodality therapy in blue, and this first outcome is adjusted metastasis-free survival. We see that in both analyses, no difference between radical cystectomy and trimodality therapy for adjusted metastasis-free survival. Again, looking at distant failure-free survival, almost exactly the same hazard ratios, 0.95 in the PSM analysis, 0.94 in the IPTW analysis. So no difference in adjusted distant failure-free survival for radical cystectomy and trimodality therapy.

Here we see again regional failure-free survival. We see slight differences, but not statistically significant between radical cystectomy and trimodality therapy. Hazard ratios of 1.56, 1.68 in both analyses, but non-statistically significant 95% confidence intervals. Disease-free survival for radical cystectomy, we see essentially the same between radical cystectomy and trimodality therapy for both types of analyses. For disease-free survival, what we do see, though, is a difference in adjusted overall survival. So again, overall survival for IPTW and propensity score matching, we can see a splitting of the curves here favoring trimodality therapy with statistically significant 95% confidence intervals favoring trimodality therapy.

Cancer-specific survival, are trending towards favoring trimodality therapy, but not statistically significant for both types of analyses compared to radical cystectomy. Then we get into some sensitivity analysis, which Rashid mentioned in the introduction. So this is patients that had radical cystectomy plus neoadjuvant or adjuvant chemo. So really the patients that got true standard of care versus trimodality therapy. We see even in these patients, there was no difference between trimodality therapy and radical cystectomy. P-value equals 0.95.

Similarly, we see no difference in adjusted cancer-specific survival. Five-year cancer-specific survival for these patients, for radical cystectomy is 80% compared to 82%. For trimodality therapy with a P-value of 0.25. Again, similar analysis, this is adjusted disease-free survival, no difference between radical cystectomy and trimodality therapy for adjusted DFS with a P-value of 0.93.

This is an important slide because this is adjusted metastasis-free survival by treatment group and center. So what this is essentially showing us is that where cystectomy was performed in this study was either Los Angeles or Toronto, as well as trimodality therapy, which is either Toronto or Boston. There was no difference in MFS by center. So this shows some of the generalizability of doing trimodality therapy at different centers. Obviously, these are centers that are centers of excellence for trimodality therapy and cystectomy, but no difference between centers, when you compare these four metrics, as we see here in this slide.

This is adjusted cancer-specific survival in patients who underwent trimodality therapy stratified by cystectomy status. There were 38 patients or 13% of patients who underwent salvage cystectomy. Majority of these were due to invasive recurrence, one was due to toxicity from trimodality therapy, and we see no difference in patient outcome for salvage cystectomy versus no salvage cystectomy. So it shows that even in patients that need to be salvaged, they could still have good outcomes among these 13% of patients.

So several important discussion points from this study. This study showed that using two statistical methods to balance groups. This multi-center study, showed no difference in metastasis-free survival, cancer-specific survival, or disease-free survival between trimodality therapy and radical cystectomy in select patients with muscle invasive-bladder cancer. This included patients who would have been eligible for both procedures. So this is not cherry picking patients. These are patients that would have been eligible for both, and outcomes for radical cystectomy and trimodality therapy were not different among centers. Importantly, the overall survival advantage for trimodality therapy was in part driven by measurable postoperative mortality patients, for patients undergoing cystectomy, which may be as high as 2.5% at 90 days compared to 0% at 90 days for trimodality therapy.

The potential role of neoadjuvant chemo with trimodality therapy requires additional study. However, there is level one evidence supporting concurrent radio desensitizing chemotherapy with trimodality therapy. Finally, trimodality therapy does require stringent diligent follow-up, as the risk of bladder recurrence is roughly 11% for muscle invasion and roughly 20% for non muscle invasive disease and requires management at centers comfortable performing post-radiotherapy cystectomies, which may be as high as 13% of the patients. We do know that muscle-invasive bladder recurrences occur in the first two to three years, whereas non-muscle invasive bladder recurrences occur even after five years, thus requiring lifelong cystoscopic evaluation.

So in conclusion, in the absence of randomized trials, which are unlikely to be carried out, this study provides the best evidence possible supporting that trimodality therapy in the setting of multidisciplinary shared decision-making should be offered to all eligible candidates with muscle-invasive bladder cancer as an oncologically equivalent alternative to cystectomy. There are ongoing investigations that are tending to further improve outcomes with trimodality therapy, including the addition of immunotherapy and certainly better predictive biomarkers for patient and treatment selection are needed.

Finally, hopefully these results will renew interest and further support for undertaking a randomized trial in this disease space. We thank you very much for your attention. We hope you enjoyed this journal club discussion of this recently published paper in Lancet Oncology.