Memorials and Mandates for Cutaneous Lymphomas

PATIENTS WITH cutaneous lymphoma present to dermatologists with malignant exaggerations of cutaneous lymphocyte biological features.¹ The puzzles presented by cutaneous T-cell lymphomas (CTCLs) are actively being pursued to understand and, one hopes, to exploit the regulation of cutaneous T cells. In the immunoregulatory repertoire of the skin are the answers to atopic dermatitis, psoriasis, contact sensitivity, and most inflammatory disorders that also present to dermatologists. The lure of understanding T-cell regulation by way of CTCL has even attracted criminal activities. One group of CTCL investigators was convicted of fraud when they uttered their version of "eureka."² Still, there are few shouts of discovery. The answers can only be achieved in a stepwise approach. The accompanying article by Kim et al³ is a fitting memorial to the extent of information available from retrospective observational studies of CTCL. The key points of that article lead us to visualize the next steps: the mandates needed to explore the impact of therapy and the goals of therapy.

Retrospective studies that have probed the severity and temperament of CTCL have come to conclusions similar to those presented in this issue of the ARCHIVES.³ Over the decades that the data for this article (or any other long-term follow-up study) were collected, the diagnosis of mycosis fungoides–type CTCL has been changing. Digitate dermatosis⁴ and lymphomatoid papulosis⁵ have been described as 2 mimics of this diagnosis. Lymphomatoid drug eruptions⁶ and even lymphomatoid keratosis⁷ have been described, most likely after many cases were diagnosed as genuine lymphoma. One of the mandates from this retrospective study by Kim et al³ is that we need minimal diagnostic criteria for patients entering a lymphoma database. Similar to the International Society for Cutaneous Lymphoma workshop on erythroderma nomenclature,⁸ a recent series of workshops are in the process of developing a set of criteria for participants to use in assessing the strength of diagnosis in patients to be reported. Certainly, the use of pathological assessment alone is flawed.⁹ The diagnostic criteria will need to include clinical, histologic, historical, and molecular factors that can be shown to be reproducible. This is not to say that a patient could only be diagnosed by a participant of these workshops. The mandate is to standardize the diagnosis for patients entering a reportable database.

Observational studies depend on the quality of data collected. If the diagnosis of mycosis fungoides is subject to change, should it also incorporate molecular techniques that are likewise subject to change? If the mandate is to provide the most precise diagnosis possible, then the diagnosis will need to in some way embrace assessments of T-cell clonality. In the current usage of TNM terminology, the technology of Sézary cell assessment is embraced despite the documented variability of this technology.¹⁰ The replacement of current technologies will undoubtedly be at the expense of improved precision, a mandate in the research of CTCL.

Once a diagnosis of cancer is made, the attention of the physician and patient is drawn to assessing its severity. The TNM staging system was developed for this purpose. The American Joint Committee on Cancer (AJCC) states that the purpose of the TNM system is to describe "the attributes of the tumor that define its behavior."^{11 (p3)} The TNM staging system for CTCL was based on observational data, and it continues to be reaffirmed by studies such as that by Kim et al.³ Staging is an important concept, but it is frequently misunderstood and misrepresented, even in the ARCHIVES. The point most commonly misunderstood about CTCL is that staging is done at diagnosis and does not change. The AJCC staging manual specifically states that staging is done at diagnosis, prior to therapy, and that "restaging" or the use of an R-stage system after a course of therapy is to be strongly discouraged.^{11 (p395)}

Recent incidents help illustrate this point. In conducting a study of stage IA CTCL, a patient was referred following an allogeneic hematopoietic stem cell transplant for CTCL. The patient had presented with stage IVA, had undergone transplantation, and was now relapsing. That is biologically different from the patient who presents de novo with a few patches of stage IA CTCL. Confusion concerning the use of stage is not limited to practice, as we can see in the ARCHIVES. Last fall, a series of 3 cases of CTCL treated with allogeneic hematopoietic stem cell transplantation were presented.¹² Case 2 was staged at IB at diagnosis, yet later the patient was restaged at IVA. Similarly, case 3 in that series was diagnosed at IIB, but after an interval of time was restaged at IVA.¹² It is important to realize that in survival studies, such as that presented by Kim et al,³ the stage of a patient's disease at diagnosis is retained throughout the course of their disease. Otherwise, there would be little decrease in the survival of patients with patch/plaque disease if they were continually reassigned to other stages should a tumor or lymph node involvement appear. In other words, a patient with stage IB disease at diagnosis dies as a patient with stage IB disease, even if that patient's body is covered with tumors at the time of death.

The tendency to use stage in an ambiguous manner is driven by the lack of any terminology or scoring system for tumor burden. T status can function as a shorthand for tumor-burden status, as could several tumor-burden assessments that have been used in recent multicenter clinical trials. Recent drug therapy trials have used a combination of target lesion scores and body surface area of involvement studies.^{13 - 14} With the abundance of studies of psoriasis therapy, dermatologists are becoming conversant with the terminology of psoriasis severity known as the PASI (Psoriasis Area and Severity Index) score.¹⁵ This modality can be readily adapted to patients with CTCL, and, given the growing familiarity with PASI scoring, it may become a more common language than T1, T2, T3, and T4. Unless a disease status terminology is proposed and verified, the ambiguous and inappropriate use of the word stage will continue. Furthermore, the limited ability of T1 through T4 to reflect disease status will also be perpetuated.

The article by Kim et al³ contributes to our understanding of severity of CTCL. In accordance with the AJCC, they have defined a key attribute of CTCL that defines its behavior. In the T-status staging system the extent of skin involvement falls into 3 categories: limited patch/plaque (T1), widespread patch plaque (T2), and greater than 80% body surface area (erythroderma, T4). The decrease in prognosis from T1 to T4 shows that CTCL is inherently bad, because the greater the involvement, the worse the outcome. That is another memorial point in this article. It is now clearly shown that increasing tumor burden leads to a decrease in expected survival. The finding that an increase in the amount of CTCL increases the severity of the disease is what underlies the approach of ablating rather than accepting this chronic eruption. However, we now need to explore the corollary that decreasing the tumor burden improves prognosis.

Staging should incorporate measures of the forces that interplay to produce lesions of CTCL. The purpose of staging is to assign prognosis on the basis of these measures. The measurable components of CTCL will change as our understanding of the disease improves. Molecular chemistry and immunohistochemistry will inevitably be components of CTCL staging. The immune status of patients is markedly variable¹⁶ and an important feature in staging¹⁷ and therapy.¹⁸ Other malignancies, such as prostate cancer and breast cancer, have had recent revisions of their TNM staging systems, incorporating molecular studies. As the pathogenesis of CTCL is revealed in a stepwise manner, each step needs to be investigated as a measurable marker of prognostic or therapeutic import.

In the article by Kim et al,³ there is a presentation of the state of the art of TNM staging in CTCL. The article is a mandate for now exploring the impact of therapy on the predicted outcomes. There are 4 definitions or concepts that cannot be made in the context of this study. Yet, the progress of this field, from observational retrospection, will be in defining these concepts. The first is to incorporate an assessment of vertical depth of involvement. The impact of this concept is shown by Kim et al in that patients with T3 disease fared worse than those with T1 disease. Such assessment has the potential to distinguish patch CTCL from plaque CTCL. Depth of involvement could also be used to establish minimal histologic criteria for a tumor. One preliminary study suggested that depth of involvement from CTCL reflects severity,¹⁹ but as a prognostic tool this would require the standard of care to include a biopsy of the thickest-appearing lesion. The criteria for diagnosis to enter a lymphoma database are under study, but there is no definition of a second critical term: relapse. Different centers use variable criteria. Without a reproducible definition of relapse, the power of oncologic terms like freedom from relapse cannot be harnessed. The third definition needed, cure, is well demonstrated in CTCL therapy studies, but definitions vary. For practical purposes, the definition of cure as being free of disease for 8 years after all therapy has ended worked in a nitrogen mustard trial²⁰ and probably should be the standard. The fourth definition needed is that for therapy, which would describe what constitutes an adequate trial of therapy. This is a step toward defining time to treatment failure, a useful oncologic concept. With these terms in hand, the mandates for future clinical research on CTCL can be realized, and, in turn, the lessons for T-cell regulation in the skin can be extrapolated to other T-cell–mediated skin diseases.

The author has no relevant financial interest in this editorial.