Spontaneous regression in chronic lymphocytic leukaemia. Clinical features of 50 cases from the ERIC registry and review of the literature

Chronic lymphocytic leukaemia (CLL) is a clinically and biologically heterogeneous disease. Beside a minority of cases with an aggressive disease at onset requiring treatment upon diagnosis, there are patients with an indolent disease who rarely require intervention. The spontaneous regression (SR) of CLL, the furthest expression of a benign clinical course, is an infrequent event whose definition criteria are heterogeneous and the pathogenic bases for which are poorly understood.1-8 To further characterize the clinical and biological features of patients with CLL experiencing SR, a dedicated international registry was initiated by ERIC (European Research Initiative on CLL) in 2012. Inclusion criteria for SR were: (i) CLL diagnosis according to the 2008 iwCLL criteria (i.e clonal B lymphocyte count at diagnosis ≥5 × 109/L with a typical immunophenotype)9; (ii) subsequent lymphocyte count decrease of 50% or more of the highest recorded value and normalization below 5 × 109/L, persisting for more than 12 months; (iii) no clinical signs/symptoms (i.e. palpable splenomegaly or lymphadenopathy) of disease during the regression phase; (iv) absence of any previous CLL treatment. Additionally, we included five patients who experienced a partial SR, defined as a sustained reduction of lymphocytosis by 50% or more. The registry enrolled cases between 2012 and 2020. In total, 50 of the 55 registered patients fulfilled the inclusion criteria, 45 with evidence of a complete (90%) and five with a partial SR (10%): 16 were from Italy, 11 from Spain, nine from Denmark, seven from the United Kingdom, three from Qatar, two from Sweden and two from India. Patients were followed for a median of 14.6 years (range 2–38). At diagnosis, the median age was 62 years (range 39–89) (Table 1, Table S1). Patients generally showed an early stage of the disease: 46 (92%) Binet A (41 Rai 0, three Rai I, one Rai II, one Rai unknown) and four (8%) B/I–II. Accordingly, most patients had normal levels of beta-2 microglobulin (23/30, 77%) and lactate dehydrogenase (LDH) (30/31, 97%), and a CD38 expression below 30% (32/36, 89%), CD49d below 30% (16/20, 80%) and ZAP70 below 20% (18/24, 75%) (Table 1, Tables S1 and S2). Fluorescence in situ hybridization (FISH) analysis, fully available in 40 patients, confirmed the low risk: 17 had no FISH lesions (42.5%), 15 had del(13q) (37.5%), six trisomy 12 (15%), one del(11q) and one del(17p) (in 30% of the cells). Sanger sequencing of the TP53 gene was performed in 16 patients and all were wild-type (WT). The immunoglobulin heavy chain variable region (IGHV) genes were sequenced in 36 patients, resulting mutated in 34/36 (94.4%), as expected. Among the 22 patients with available information, the median percentage of identity was 93.8%, ranging from 64.1% to 96.9%. Regarding the IGHV repertoire, available in 29 cases, 15 (52%) cases showed the usage of the VH3 family, 11 (38%) of VH4 and three of VH1 (10%). Recurrent IGHV usage was shown in four (14%) patients with VH3-23, four (14%) with VH4-34 and three (10%) with VH3-30 (Table 1, Table S1). The IGHV repertoire of 29 patients from the present series and, when analysed, of SR from the literature4, 5, 10 was compared with 2424 Rai 0 CLL patients belonging to the general ERIC database (data not shown).11 After adjustment for multiple testing (Bonferroni correction), the IGHV family usage and gene repertoire were not significantly different between the two groups, although we found an enrichment of the VH3-15 and the VH3-30 genes among SR patients (9% and 13% respectively) versus the control cohort (3% and 6% respectively). No difference in the IGHV usage was found in a comparison with an extended series of 84 ultrastable CLL, who did not require treatment for at least 10 years from diagnosis.12 del(13q): 15/40 (37.5%) Normal: 17/40 (42.5%) Trisomy 12: 6/40 (15%) del(11q): 1/41 (2.4%) del(17p): 1/42 (2.4%) 2 VH1-2; 1 VH1-46 (n = 3) 2 VH3-15; 2 VH3-21; 4 VH3-23; 3 VH3-30; 2 VH3-33; 1 VH3-34; 1 VH3-43 (n = 15) 1 VH4-4; 1 VH4-30; 1 VH4-31; 4 VH4-34; 2 VH4-59; 2 VH4-61 (n = 11) 2 viral respiratory infections; 1 pneumonia; 1 bacterial infection 1 intracerebral haemorrhage; 1 cerebral stroke 1 pelvis fracture 3 COPD 1 chronic hepatitis C 1 rheumatoid arthritis; 1 polyarthritis 6 neoplasias (2 skin, bladder, colon, in-situ cervix, Richter syndrome) concomitant to SR 6 neoplasias (2 breast, 2 prostate, GIST, stomach) prior to SR 3 neoplasias (breast, myelodysplasia, ovary) subsequent to the SR The median lymphocyte count at diagnosis was 11.6 × 109/L (range 5.1–51.9). After diagnosis, all patients were observed for their disease course. A median lymphocyte peak of 15.6 × 109/L (range 6.3–107) was recorded after a median follow-up of 3.5 years (range 0.3–18.6) (Figure 1). In the 45 patients with a complete SR, after a median of 8.1 years from diagnosis (range 0.6–25), the median lymphocyte count at registration was 3.4 × 109/L (0.9–5.0). At the last observation time, lymphocytes were 2.8 × 109/L (0.5–4.9), with 0.5 × 109/L CLL cells (0–3.6) in the 11 cases who were evaluated by flow cytometry. Only in one case, flow cytometry documented the absence of circulating CLL cells, up to the last follow-up three years later. Of the four patients in stage B/II at diagnosis, two converted to A/0 and one to A/I. In the five patients with a partial SR, the median lymphocyte count at regression was 6.3 × 109/L, with a median lymphocyte decrease of 75.1% (69–89%), and at the last follow-up a median lymphocyte count of 5.6 × 109/L (5.3–11.8) (Figure 1). In 23 patients (46%) with a complete SR, several clinical conditions were present at the time of CLL regression, mostly infections or second neoplasias, and two cases presented autoimmune disorders (rheumatoid arthritis, polyarthritis) (Table 1, Table S1). More in detail, 12 patients developed 15 second neoplasias that were detected prior (two breast, two prostate, gastrointestinal stromal tumour, stomach), concomitantly [two skin cancer, Richter's syndrome (clonally unrelated, Table S1), colon, in situ cervix, bladder] and subsequently (breast, myelodysplastic syndrome, ovary) to SR. No patient but one underwent treatment for a second neoplasm before SR. No clear correlation with CLL-unrelated medication and SR was observed. In March 2022, an update of the follow-up of the patients included in this study was obtained for all but five cases. Of the 40 cases in SR, all remain in SR at a median time of 14.7 years (range 3.4–38) from diagnosis, and with a median of 5.52 years (range 1.2–17.5) follow-up from the achievement of a SR. Of the five patients initially categorized as in partial SR, three have subsequently converted to SR, one remains in partial SR and in one patient who achieved a partial SR after 19 years from diagnosis the lymphocyte count started to increase five years later. Thirty-seven patients were alive at the last follow-up, 13 had died (two dementia-related causes, one pneumonia, one warfarin-related bleeding, one ovarian cancer, eight because of unknown causes); no patient died due to CLL progression. We herein document that a small proportion of patients with CLL may undergo a SR of their disease, with an estimated frequency of 0.2%–1%. SR occurs preferentially in patients with low-risk disease, i.e. early Binet or Rai stage, low blood lymphocyte count, normal beta-2 microglobulin and LDH, negativity of CD38/CD49d, TP53 WT, mutated IGHV, favourable FISH in more than 80% of cases, albeit cases of SR in CLL patients displaying trisomy 12, del(11q) and del(17p) have been reported (Table 1, Tables S1 and S2). In the ERIC registry, the occurrence of SR was observed mostly after a prolonged disease history, with a median time of about eight years and up to more than 20 years from diagnosis. The SR duration ranged between one and 17 years (Table S1) and no relapse has so far been recorded. SR can be correlated with clinical conditions, present at the time of the CLL regression in 46% of cases, such as infections, second neoplasias or, less frequently, autoimmune disorders. Fifteen second neoplasias, present in 12 patients out of 50 SR (24%), were documented prior in six, concomitantly in six and subsequently to the SR in three. Interestingly, the regression of the disease may be complete, although with a small leukaemic clone still detectable by flow cytometry in almost all cases. In our series with a long follow-up, one out of 45 complete SR (2%) showed an undetectable CLL in the peripheral blood. To the best of our knowledge, the present ERIC series of CLL patients undergoing SR is the largest so far reported. A review of the literature on SR in CLL is provided in Material S1 and Table S3. Several issues remain open, such as why some patients with CLL, but not others, develop a SR, making the ultimate basis of this phenomenon elusive. Given its rarity — that could offer important insights into the immunosurveillance mechanisms of CLL and thereby into potential treatment strategies — the comprehensive study of SR based on uniform criteria and a careful follow-up as in the registry launched by ERIC is warranted. Ilaria Del Giudice designed the study, managed the registry and wrote the paper; Luca Vincenzo Cappelli performed the statistical analysis; Julio Delgado, Carsten Utoft Niemann, Michael Asger Andersen, Emelie Hamotal Curovic Rotbain, Kathrine Aarup, Renata Walewska, Andrea Visentin, Marina Deodato, Anna Maria Frustaci, Chiara Cavalloni, Massimo Gentile, Mohamed A. Yassin, Deepesh Lad, Lydia Scarfò, Max Flogegard and Mattias Mattsson provided patients with SR to the registry and updated the follow-up; Sara Raponi, Caterina Ilari and Irene Della Starza performed the IGHV mutations of ultrastable CLL, CLL with SR and the reported unselected CLL of Haematology Sapienza University; Ester M. Orlandi, Alessandra Tedeschi, Livio Trentin and Gianpietro Semenzato provided patients with SR to the registry; Anna Guarini critically analysed the results and revised the paper; Emili Montserrat and Robin Foà critically revised the manuscript; Paolo Ghia, Emili Montserrat and Robin Foà provided support to the initiative of the registry within ERIC. The authors wish to thank the Associazione Italiana per la Ricerca sul Cancro (AIRC), 5×1000 Metastases Special Program, No. 21198, Milan, Italy (Robin Foà) and the ERIC group (ERIC website: http://www.ericll.org/; ERIC registry of SR: http://www.ericll.org/projects/), particularly Kostas Stamatopoulos and Andreas Agathangelidis. The authors wish to dedicate this manuscript to the memory of Daniel Catovsky, for his pivotal contribution to the knowledge of the clinical and biological features of CLL, including SR. Ilaria Del Giudice — AbbVie, AstraZeneca, Janssen outside the submitted work. Carsten Utoft Niemann — funding and/or consultancy fees from AstraZeneca, Abbvie, Janssen, Roche, CSL Behring, Lily, Takeda, Octapharma, Beigene and Genmab outside the submitted work. Emelie Hamotal Curovic Rotbain — travel grants and/or consultancy fees from AstraZeneca, Abbvie and Janssen, outside this study. Lydia Scarfò — honoraria from AbbVie, AstraZeneca, Beigene and Janssen outside the submitted work. Alessandra Tedeschi — advisory board Janssen, Abbvie, Beigene and AstraZeneca. Gianpietro Semenzato — speaker, advisory board, consultancy fees and/or research grants from Novartis, Takeda, Roche and Janssen-Cilag outside the submitted work. Robin Foà — speaker bureau for Amgen, Novartis and AstraZeneca outside the submitted work. Data S1. Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.