Order out of chaos

In this issue of Blood, Inamoto et al show that chronic graft-versus-host disease (cGVHD) has emerged from the chaos as a distinct disease with validated staging and response criteria. Inomata et al also show that there is order in steroid refractory cGVHD that can be exploited to better design and interpret therapeutic trials.¹ 

Basic and clinical interest in chronic graft-versus-host disease (cGVHD) has exploded since the publication of the National Institutes of Health (NIH) Consensus Criteria for cGVHD in 2005-2006 (summarized in Pavletic et al² ). Why this explosion? Quite simply, communication. Instead of each investigator living in his or her own parallel universe and having no effective way of communicating the extent of disease or the response to treatment, investigators now have a common language they can use to work with each other. Admittedly, much of the initial published work has been in validating and refining the NIH criteria.^3,4  But, as demonstrated by the Inamoto et al article, the order necessary for true basic and clinical advances is becoming evident.

Inamoto et al report a retrospective study of 312 patients who received second-line systemic treatment for cGVHD to characterize causes of treatment failure, to identify prognostic factors associated with treatment failure, and to develop shorter-term end points for trials that test second-line systemic treatment of cGVHD.¹  The primary end point was failure-free survival, defined by the absence of third-line treatment, nonrelapse mortality, and recurrent malignancy during second-line treatment. Treatment change was the overwhelming cause of treatment failure, as shown in the figure. By 6 months, 34% of patients were on third-line treatment, and an additional 10% died or had relapsed. Multivariate analysis showed two expected (lower gastrointestinal involvement at second-line treatment and severe NIH global score) and one unexpected (high-risk disease at transplantation) risk factors for treatment failure. These three factors were used to define risk groups, and success rates at 6 months were calculated for each risk group (without or with various steroid dose limits).

Many have been frustrated by the lack of reproducibility of trial results in steroid refractory cGVHD. This same group of investigators has championed the premise that the problem is due to lack rigor in study design; it is hard to interpret, compare, or reproduce studies without clarity in the study design and assumptions.⁵  The current study suggests that at least some of the splay in results may be due to the risk factors of patients enrolled in the initial vs subsequent confirmatory studies. Although most would have predicted that those with severe involvement of multiple organs and possibly lower gastrointestinal involvement would have a worse treatment outcome, few would have considered the underlying disease indication for transplant as a risk factor or that response to treatment based on these three risk factors could be grouped in the manner demonstrated here. It would be intriguing to reexamine the results of steroid refractory cGVHD trials in this light. However, most steroid refractory cGVHD trials were not limited to second-line treatment, and multiple studies have suggested that responses were less common in heavily pretreated patients. Hopefully, Inamoto and his collaborators will extend this study to look at those who did require third-line and beyond lines of treatment and to look at those patients with “stable” disease. The tempo of response needs to be better defined in order to better understand when stable truly indicates failure. The one significant caution with the data presented in this article is that “lack of improvement after at least 2 weeks of initial treatment” may be too small a window to declare a treatment failure, especially in patients with sclerotic skin changes.

The investigators also propose that the combination of steroid dose (which correlated with the subsequent withdrawal of immunosuppressive treatment) and failure-free survival rates at 6 months could be used as the basis for a clinically relevant and efficient shorter-term end point in studies evaluating second-line systemic treatment of cGVHD. The protracted time needed to conduct cGVHD trials has been a substantial obstacle to progress in clinical care of this disorder. Neither drug companies nor junior investigators trying to establish their careers are interested in trials requiring 5 to 10 years to complete. Patients deserve better than an educated guess as to their best option for treatment, especially given that half will be placed on second-line treatment within 1 year of diagnosis.^6,7  Although enrolled patients would continue to need to be followed, these data suggest that a 6-month end point should give a good indication of the promise (or lack or promise) of the second-line treatment.

There has been a similar, but more modest, explosion in the basic science investigation of cGVHD. It is still hard to see order in the data, but at some point soon, the bigger picture of the pathophysiologic mechanism producing cGVHD will become clearer. As recently summarized by Paczesny, biomarkers for cGVHD should be pursued and will help drive this understanding.⁸  When these become available, the synergy of clinical and laboratory end points should allow for even more rapid clinical evaluation of new and better options for first-line and subsequent treatment. Because it is likely that multiple immunologic pathways are involved, it will also allow patients to receive more targeted treatment. The study by Inamoto et al should bring us all closer to the secret order of cGVHD, a day that will be celebrated by transplantation centers and patients alike.