Incidence, Pattern and Clinical Correlation of Bone Marrow Involvement In Adult T-Cell Leukaemia/Lymphoma (ATLL)

Adult T-cell leukaemia/lymphoma (ATLL) is an aggressive T-cell lymphoma associated with the retrovirus human T-cell lymphotropic virus type 1 (HTLV-1). Cases are classified into one of four subtypes – acute, lymphomatous, chronic and smouldering - according to the criteria of Shimoyama. Essential data necessary for accurate classification are the absolute lymphocyte count, percentage of abnormal lymphocytes, corrected calcium, lactate dehydrogenase measurement and pattern of tissue involvement but assessment of bone marrow (BM) is not essential for classification. Reported median survival in aggressive ATLL is 6.2 months for acute and 10.2 months for lymphomatous subtypes. Data regarding BM involvement in ATLL is scarce coming from small series or case reports and ATLL subtype is rarely reported. In one study BM involvement occurred in 35% of patients with aggressive ATLL, subtype not specified, and was an independent predictor of poor overall survival (Takatsuki et al,2007).

To determine the frequency, pattern and morphological features of BM involvement in ATLL.

A retrospective analysis of pathology records of patients with ATLL diagnosed between 2000 and 2009 in whom bone marrow staging by trephine examination was undertaken. In addition to morphological analysis immunohistochemistry (IHC) was performed using a panel comprising monoclonal antibodies against CD2/CD3/CD4/CD5/CD7/CD8/CD20/CD25/CD30/CD56. CD25 staining was further analysed as membranous, cytoplasmic and golgi localisation. HTLV-1 infection was confirmed serologically in all cases.

Of the 19 cases 10 had lymphomatous, 8 acute and 1 chronic ATLL. The median age was 49.4 years (range 31.2–74.5) and 11 cases were female. BM involvement was detected in 12 cases (63%). This was evident morphologically in 10 patients, including all 9 with leukemic ATLL and 1 with lymphomatous disease. Disease bulk ranged from 5–40% overall but was <20% in 4/8 with acute ATLL. In 2 additional patients, both ATLL lymphoma, BM involvement was confirmed on IHC only but was minimal (<1%). The absolute lymphocyte count was significantly higher in cases with BM involvement (10.2×10⁹/L v 1.2×10⁹/L, p=0.004) but there was no correlation between degree of BM involvement and lymphocyte count. There was no significant difference in haemoglobin, platelet count, corrected calcium or lactate dehydrogenase levels. The pattern of infiltration was a mixture of interstitial (80%) and nodular (90%) with juxta-trabecular (30%) and peri-sinusoidal (10%) infiltrates also seen. Cell size varied from small to large with large lymphocytes more frequent in acute ATLL (75% of cases). Nuclear irregularity was present in 3/10, all of whom had a range of small, medium and large lymphocytes present. BM eosinophilia was present in 4/19 cases, all with acute ATLL. Of the 5 patients with hypercalcaemia at presentation, all with acute ATLL, bone resorption was observed in 2/5. On IHC the morphologically abnormal cells were CD3+ (11/12), CD2+ (9/10) and CD4+ (9/10). One case was CD4/CD8 dual positive. CD25 was positive in 100% and staining was a mixture of membranous (100%), cytoplasmic (89%) and golgi (22%). CD7 and CD56 were negative in all cases. CD5 expression was lost in 2/8 cases. CD30 was positive on large cells only and this was noted in 2/7 cases. Overall median survival was 8.5 months (range 0.4–72.5) and outcome did not differ according to the presence of bone marrow disease (BM+ 8.0 v BM- 10.4 months, p=0.33).

BM involvement was present in all cases with leukaemic ATLL but only in a minority with lymphomatous ATLL, in 2 cases detectable by IHC only; the latter suggests the usefulness of this technique in residual disease assessment post therapy. Appearances were heterogenous with nodular and interstitial infiltrates frequently present in the same case. Eosinophilia and bone resorption were present in cases of acute ATLL only. The significance of BM involvement in ATLL prognosis remains uncertain.

No relevant conflicts of interest to declare.