By BISHAL GYAWALI MD, PhD
Hey, I’m back!
Well, you might not have noticed that my blogs were missing for the last three months but anyways, its good to be back. I was having a little time off blogs and social media as I was transitioning in my career but now I am back. Sometimes, it is very difficult to manage time for things that you must do versus things you enjoy doing, especially when these two don’t intersect. For me, these last few months the things I had to do were all bureaucratic while I couldn’t find the time for things I enjoy doing like writing these blogs. But now that we are back, let’s recap what has happened in the oncology world in the year 2019 so far. I can’t cover all of them, but will try to summarise the major events in oncology.
Hundred Foxes’ Howl versus One LION’s Roar
In my country, there is a saying that goes somewhat like the roar of one lion will scare hundreds of howling foxes away. In medicine, I guess, it translates as one good RCT trumps the results from hundreds of observational studies. For patients with advanced ovarian cancer, primary surgery to achieve complete resection is the most important treatment and prognostic factor. However, what to do with the lymph nodes is a question that has troubled the oncology community for a long time. Logically, it makes sense to remove the lymph nodes too because they are the sanctuary sites for cancer cells. However, lymph node dissection carries high morbidity. Although multiple observational studies suggested a survival benefit with lymph node dissection, the LION trial, now published in the NEJM, shows that for women with macroscopic complete resection of primary tumour, lymph node dissection increases morbidity (postoperative complications) and post-operative mortality rates but doesn’t improve survival. I am glad that this trial was carried out and these results will now save many women with ovarian cancer worldwide from unnecessary harmful procedures, but I am also sad that we didn’t answer this question until now and thus, many patients suffered unnecessarily. I hope this LION’s roar scares us from jumping to conclusions based on logic or observational data alone and without RCT evidence in future. Another lesson here is the importance of public funds in supporting RCTs like these.
Practical exercises on trial methodologies and reporting
Going through oncology articles published in top journals in the last couple of months seemed more like a practical course on spotting methodological issues in publications. I don’t have the time to discuss them all here but for any student of research or trial methodologies, these papers will be good learning exercises. I have tried to focus more on statistical than clinical aspects of these trials because they are in general not practice changing yet.
First, in this phase 1/2 trial, a drug in competition for the most difficult drug to pronounce/spell in oncology, sacituzumabgovitecan-hziy, was tested in patients with refractory metastatic triple negative breast cancer. Although the responses look impressive, there is no mention of the criteria for success. Shouldn’t there be an a-priori definition of when success can be claimed for a phase 2? How do we decide when to take them to phase 3?
Second, these two RCTs of direct oral anticoagulants to prevent VTE in ambulatory cancer patients are an exercise on understanding intention-to-treat versus per-protocol treatment. Another lesson here is to look at the absolute difference in event rates besides the hazard ratio. In any case, the burden of therapy here seems to outweigh the benefits.
Third, in this RCT of TDM1 as adjuvant therapy for patients with HER2 positive breast cancer who had received a trastuzumab-containing neoadjuvant regimen but had residual invasive breast cancer, the conclusion reads “the risk of recurrence of invasive breast cancer or death was 50 percent lower with adjuvant T-DM1 than with trastuzumab alone”. The hazard ratio is 0.5 but hazard ratio is not the same as risk ratio. This is a common misconception as we have previously shown in this experiment. The difference in absolute percentage of patients who remained free of invasive disease at 3 years was 11 percent (88 percent v 77 percent) which is impressive, but it’s not 50 percent. Finally, with a median of more than 40 months of follow-up, the OS hasn’t seen significant improvement. Also a tricky question now is whether TDM1 retains the survival benefits when the disease has relapsed now that patients have already used it in the adjuvant setting.
Fourth, this RCT of ramucirumab after sorafenib in hepatocellular cancer provides multiple lessons: Clinically meaningful versus statistically significant difference in outcomes (OS benefit of only 1.2 months but p = 0.0199), highly selected patient population for enrollment (the control arm OS of over 7 months in second-line hepatocellular cancer) and the use of this sentence in the conclusion despite 3 fatal adverse events and an increased percentage of serious adverse events within the ramucirumab arm: “Ramucirumab was well tolerated, with a manageable safety profile.” If you don’t know why I am not happy with this statement, please read this article we published in the BMJ that addresses this exact point.
Another drug that improves an endpoint that you didn’t even know existed until few years ago
Darolutamide has shown to improve metastasis-free survival in patients with non-metastatic castration-resistant prostate cancer in the ARAMIS trial. Metastasis-free survival is a new surrogate endpoint, about which I have discussed in detail in an earlier blog. To be fair, unlike enzalutamide, darolutamide has also shown improvements in overall survival but only 15 percent of the patients in placebo arm received enzalutamide subsequently. When this trial was presented at ASCO GU 19, the discussant Prof. Ian Davis’s summary slide which nicely highlighted all the caveats floated around Twitter, so I will just put his slide here: https://twitter.com/birensaraiya/status/1096175780579561472.
Also, while we are talking about prostate cancer, I’d like to highlight this study which showed that the combination of radium plus abiraterone was harmful than abiraterone alone. Another example to keep in mind to remind ourselves to exercise caution in recommending A B when both A and B are approved agents for a given cancer.
We need more publicly funded trials
An important trial in glioblastoma was published in the Lancet in February. Until now, the Stupp regimen has remained the standard of care but this new RCT of Stupp regimen plus lomustine showed a significant improvement in OS of 17 months compared to Stupp regimen alone without lomustine. The sample size was small and the authors conclude the abstract as “Our results suggest that lomustine-temozolomide chemotherapy might improve survival compared with temozolomide standard therapy in patients with newly diagnosed glioblastoma with methylated MGMT promoter. The findings should be interpreted with caution, owing to the small size of the trial.” I totally appreciate the honest interpretation from the authors here but also can’t help wonder how this paragraph might have read if this was the trial of immunotherapy instead of lomustine and funded by industry rather than the German Federal Ministry of Education and Research.
Let me take a selfie
After a long time, Vinay Prasad and I have written a paper together. This time we talk about risk-benefit trade-offs in the adjuvant treatment setting, especially in the absence of robust data, because the threshold for treatments are different from the metastatic setting. We also list various examples of drugs that are effective in metastatic setting but failed as adjuvant therapy.
Dr. Gyawali is a research fellow at Program On Regulation, Therapeutics And Law (PORTAL) at Brigham and Women’s Hospital/Harvard Medical School. The opinions expressed herein are his own. This post originally appeared on ecancer here.
Is there an underlying bias regarding “significance” during pre-publication processes? If so, how does it sustain itself?
thank you for share.