More is more or less is more?

There have been many advances in uro-oncology lately. In January, I dedicated a blog entirely to uro-oncology, and genitourinary cancers continue to dominate the oncology news in February. Let’s start with the big apple of discord in oncology: treatment of recurrent prostate cancer. RTOG 9601 showed benefit in overall survival by the addition of 24 months of antiandrogen therapy to radiation at the time of recurrence of prostate cancer after surgery. A key inclusion criterion was a life expectancy of more than 10 years.  Around 24% and 1% of patients enrolled were aged between 70-79 and > 80 years respectively. It is interesting how patients >80 years old were considered to have a life expectancy of >10 years because, irrespective of prostate cancer, no country has an average life expectancy of over 90 for males. Furthermore, outcomes data on age-specific subgroups have not been presented. I would like to see if patients of age 70-79 and >80 derived the same benefit as the rest. This is important in considering treatment choices for elderly patients whose priorities and risk-benefit threshold for tradeoffs are different. Other data to consider in making treatment decisions is the incidence of gynecomastia (70% in antiandrogen group vs 11% in placebo group).  Furthermore, the survival curves don’t separate until 3 years after randomization. Indeed, for patients with low life expectancy or low risk cancers, observation alone (not even radiotherapy) at the time of recurrence is still an acceptable strategy.

Just appropriate is more

The famous HERA trial that established adjuvant trastuzumab as the standard of care in HER2 positive breast cancer is back with 11 year results. Luckily, the conclusions stay the same: 1 year of trastuzumab is better than no trastuzumab or 2 years of trastuzumab. Neither less nor more is more, the right balance is the key- something Buddha expounded on more than 2500 years ago!  However, I was intrigued to see the actual difference: 10 year DFS was 63% in the observation alone group versus 69% in both the 1 year and 2 year trastuzumab groups. The absolute gain in absolute survival at 12 years with 1 year of trastuzumab versus observation was 6.5%. Trastuzumab is a classic example of precision medicine by targeting HER2. But these data tell me that while some patients definitely benefit from adjuvant trastuzumab, many don’t. The problem is that we are drugging all patients because we don’t know which minority will benefit. Where’s precision in all this? We desperately need better biomarkers, otherwise we might as well name it “uncertain oncology” which actually is closer to the truth.

What about hairs on your head? More is more?

Alopecia is an important side effect of many cytotoxic drugs. This side effect is particularly concerning to female patients. Unfortunately, the standard cytotoxic drugs used in the treatment of breast cancer such as anthracyclines and taxanes are notorious for causing alopecia. Hence, some patients avoid chemo altogether or select alternative drugs for the fear of alopecia. Unfortunately, preventive measures against alopecia are limited and RCT evidence lacking. Thus, it is welcome that two prospective studies in JAMA have established that the use of a scalp cooling device protects against severe chemo-induced alopecia (>50% fall) in nearly 50% of patients. Although one was a prospective cohort study with a control of very small sample size, the other was an RCT with a rare appropriate acronym SCALP. The SCALP trial provides comfort to many patients and eases their apprehension of starting chemotherapy. However, one important finding was that although the device helped in preventing severe alopecia, it was not associated with improvements in quality of life. An accompanying editorial argues that “the overall effects of the breast cancer diagnosis, surgery and chemotherapy may have had a substantial influence on the patient-reported outcomes in both groups. This could diminish the likelihood of detecting a difference in quality of life related to a lower rate of alopecia.” But isn’t that the point? Aren’t we supposed to manage the whole patient together and not individual symptoms separately? If other disease related events stress the patient so much that alopecia is no longer a concern, should we necessarily actively prevent alopecia? Don’t get me wrong - I have seen many patients with substantial fear of chemotherapy due to alopecia. My point is that we are not supposed to generalise our patients. Isn’t that the best way to practice personalised medicine? Addressing each patient’s concern separately. So, for patients who put alopecia high in their list of concerns, the scalp cooling device seems to be a good device but proper discussion of risk-benefit must precede the decision. The risk in this case is usually financial toxicity. According to the paper, this device costs between US$1500-3000 and is not covered by insurance. This should also be balanced against the fact that this device doesn’t guarantee hair preservation - it prevents >50% hair loss in nearly 50% patients. Not complete hair preservation, and not for all patients.

Increasing options in urothelial cancer

Options in the second line treatment of urothelial cancer are limited. However, the landscape is slowly changing with the introduction of immunotherapies. While atezolizumab and nivolumab are already approved based on uncontrolled phase 2 data, this phase 3 trial showed improved survival with pembrolizumab versus chemotherapy. Thus, pembrolizumab should be the standard treatment in second line therapy after patients progress on first line platinum based chemotherapy. The benefit in median survival is nearly 3 months - however, median doesn’t seem to capture the benefit of immunotherapies well. The median PFS was in fact shorter with pembrolizumab.

Now we are faced with some clinical dilemmas. Firstly, which immunotherapy to prefer in 2^nd line? The clear answer is pembrolizumab, because it is the only one with proven phase 3 survival benefit. Second, will immunotherapy work in first-line? Hopefully, but we don’t have data yet to support that practice. Third, is it ok to use one immunotherapy after progression on the other? Absolutely not at this moment. All these drugs inhibit the PD1-PD-L1 interaction and there’s no reason to believe that one would work when the others have failed.  Fourth, how long should pembrolizumab be given? In this study, it was given until RECIST based disease progression or until 2 years. It’s clear from the PFS and OS curves that RECIST based disease progression doesn’t predict efficacy. The PFS curves start to favour pembrolizumab after 6 months. I would assume that giving pembrolizumab for 6 months might be enough but this is only my opinion, not evidence. What should we do then? Remember the non-inferiority trial of zoledronic acid I praised last month? Yes, we need a non-inferiority trial of a shorter duration of immunotherapy versus using such expensive agents until disease progression, especially because immunotherapies are known to produce lasting responses and continued responses even after stopping the drug. We owe such trials to our patients (to reduce unnecessary toxicity) and society (to reduce financial toxicity overall). Fifth, do all patients need pembrolizumab? No. Not all patients benefit from pembrolizumab and unfortunately, some patients die quite early. Thus, it’s not a miracle drug and we need good biomarkers to identify patients who benefit from those who don’t. Finally, is PD-L1 an appropriate biomarker? No! It has not been a good predictive marker of immunotherapy for any tumour type. We need to start looking beyond PD-L1. 

Putting yourself in the patient’s shoes

This editorial by a statistician has deeply touched me. Commenting on a network meta-analysis, Helena Kraemer writes: “There are two tasks that statistical reviewers of clinical research articles cannot do. First, statisticians cannot fully appreciate the clinical or research importance of the issue under study, because they generally lack the expertise of either a clinician or a clinical researcher. However, it is certainly possible for statisticians to imagine themselves as patients whose well-being may depend on the veracity of the conclusions presented.” This made me wonder how many clinicians actually making those life and death decisions imagine themselves as patients while deciding on particular treatment for the cancer patients in their clinic? Are we oncologists well conversant with the evidence and appropriateness of that evidence to the patient? I found it very concerning that while statisticians who are not directly treating patients are so serious about the hazards of drawing inappropriate conclusions based on statistics alone, some of the clinicians or clinical researchers casually treat the data and draw inappropriate conclusions that can potentially have a direct impact on real people's lives.

Other mentions

1. I have written earlier about the lack of evidence to support upfront combination immune checkpoint blockade. This strategy has never been tested against sequential blockade. To add more weight to the argument, a recent analysis has shown that sequential checkpoint blockade is more cost-effective than upfront combination.
2. In a confirmatory phase 3 trial, osimertinib prolonged PFS versus chemotherapy among patients with EGFR mutation positive non-small cell lung cancer who had progressed on first line EGFR TKI due to T790M resistance mutations. Osimertinib has already been approved in this setting but this confirmatory phase3 trial gives key insights. The PFS hazard ratio is good (0.30) and CNS activity encouraging. However, overall survival and quality of life data - things that matter the most in advanced cancer patients - are not available yet.
3. There is a collection of some very nice papers (although I don’t agree with all of them) related to contemporary oncology in the February 9 issue of Cell.
4. Finally, don’t trust claims made from subgroup analyses as this study has shown.

Bishal Gyawali, MD is a postgraduate trainee in medical oncology at the Graduate School of Medicine, Nagoya University, Japan, where he is also a PhD candidate under the Japanese government scholarship. His areas of clinical and research interests include evidence-based oncology practice, global oncology, cancer policy, cancer management in resource-limited settings, financial toxicities of cancer treatment, clinical trial methods and supportive treatment of cancer. He has no conflicts of interest to declare. Dr Gyawali tweets at @oncology_bg. Read his previous blog posts here.