A highly practical and interesting study, CARD, was recently presented at ESMO and published in the NEJM. It’s a study that answers a lot of questions, creates a few others, and can be translated into the clinic relatively quickly.

The CARD study randomized patients with castration-resistant prostate cancer (CRPC) to either treatment with cabazitaxel  ( taxane chemotherapy) or a second ‘sequence’ of androgen receptor (AR) targeted therapy (ARTT) – enzalutamide in patients with prior abiraterone exposure, or vice versa.

Patients were allowed to have received prior docetaxel as is the standard of care in many parts of the world for high volume castration sensitive metastatic disease. All patients were required to have received a prior ARTT, and patients with prior docetaxel treatment in metastatic castration-resistant prostate cancer (mCRPC) were enrolled.  Patients were stratified by duration of therapy with the prior ARTT in order to avoid an imbalance of responding versus non-responding patients.

Here are my thoughts on what this study showed us, and maybe didn’t’ show us.

  1. More evidence that sequencing AR targeted drugs is not the way to go.

Thanks to the work of Kim Chi in Vancouver and Gerhardt Attard in London, who have done the tough work of conducting prospective clinical trials in which patients were methodically switched from enzalutamide to abiraterone or abiraterone to enzalutamide, demonstrating very poor rates of response, we already have a growing body of evidence that ‘sequencing’ is not really something we should be doing. The CARD trial layers in chemotherapy to this equation and adds to prior retrospective series suggesting that the response rate to taxane chemotherapy is essentially unperturbed by prior ARTT exposure.

The CARD trial l required prior ARTT and progression on this therapy. Interestingly, in this patient population, the duration of prior ARTT therapy was only a median of about 8 months. This is quite a bit shorter than the 16.5 and 18 months that was reported with abiraterone in COU-302 and enzalutamide in PREVAIL, respectively. It is more in line with the results of the COU301 and AFFIRM study populations, all of whom were post-docetaxel. So, even though stratification took place it almost seems like there was a selection for patients with relatively high risk and ARTT resistant disease. That’s a theme overall that emerges from the CARD study: these were relatively sick patients with aggressive disease, not the asymptomatic or mildly symptomatic patients of the aforementioned phase III’s.

Also keep in mind that, in CARD, 61% of patients received their ARTT after docetaxel, and that in 90% of patients, the prior docetaxel was administered for mCRPC, not hormone-sensitive prostate cancer (HSPC). CARD is, in actuality, a study of third-line therapy and in most cases, the cabazitaxel is being delivered ‘on label’ in the post docetaxel setting.

  1. These data are highly relevant for the symptomatic patient. Less so for the asymptomatic?

A surprising feature for me was the observation that two-thirds of the patients enrolled in the CARD trial had disease-related pain at the start of the study, and that worsening pain was a key feature of their pre-enrollment disease progression. Of the three kinds of disease progression in mCRPC ( PSA only, PSA with new radiographic lesions, PSA and Pain progression) the latter is clearly worse. We have shown in prior analyses that such patients who are experiencing unequivocal clinical progression, have a shortened survival compared to those experiencing disease progression by one of the other definitions. Thus CARD is a test of taxanes in a largely symptomatic, heavily pretreated population – not too dissimilar from the original TROPIC study that led to the approval of cabazitaxel.

  1. Taxane exposure is not Taxane resistance?

I’m not sure we can answer this one from the CARD trial because only 10-14% of the patients in the Cabazitaxel or A/E arms, respectively, in the metastatic hormone-sensitive prostate cancer (mHSPC) setting. Yet, if we are using taxanes in the HSPC setting it is important to remember that in that setting patients are generally treated for a finite duration of six cycles rather than until disease progression. Following this they are observed. The fact that they are not treated to disease progression on their first taxane is important. CARD does not really answer the question of how effective taxanes are in the post mHSPC Taxane-treated setting as the numbers are too small. The CARD study used cabazitaxel at a dose of 25 mg/m2 with GCSF support in all patients. The results of the study constitute level one evidence that taxane chemotherapy – this drug at this dose, is superior to a second ARTT.

Finally, I will indicate that I hope this is not the last we hear from the CARD investigators and I hope we will see more molecular subtyping in the near future. While it is interesting and important to tease out the patient subtypes based on age, Gleason grade  and pain status as we interpret new data such as this, it is critically important, as we move our field forward, to describe the molecular phenotypes we are observing. For example, in the mCRPC setting it has been shown that TP53 alteration substantially alters the prognosis of patients, and I would like to know if such patients were evenly distributed across this study and how well cabazitaxel performed in the TP53 altered setting. Furthermore, a very nice recent publication from Wassim Abida at Memorial Sloan Kettering Cancer Center revealed that many patients with progressive disease on ARTT harbor a form of the disease in which RB1 has been lost, and that RB1 loss is associated with a substantially worse overall survival. Yet, we don’t have any real prospective information on the efficacy of chemotherapy in that patient population, or the potential efficacy or ARTT in patients with versus without intact RB1.

I guess I’m getting a little impatient. For me, and I hope for the field in general, I hope to see a more broad expression and appreciation of tumor subtypes in clinical trial reporting. Of course, we want molecular testing to guide therapy, and It can from time to time ( eg in BRCA2 mutation, or MSI high status) but I would also find it informative to see the distribution of patients on this ( or any other trial ) with RB1 loss, TP53 loss, PTEN loss etc. We still report the Gleason score of the prostate in patients with mCRPC. It tells us something and is useful. I just think we can do better by digging a little deeper and focusing on the molecular pathology of the tumor we are treating, not the one the was biopsied or removed sometimes years earlier.

Written by: Charles J. Ryan, MD is the B.J. Kennedy Chair in Clinical Medical Oncology at the University of Minnesota and Director of the Division of Hematology, Oncology and Transplantation.