Bone Involvement On Pet/Ct Predicts Event-Free Survival In Follicular Lymphoma Grade 3b

Follicular lymphoma (FL) Grade 3B (FL3B) is a subtype of FL with a distinct cytomorphological, immunohistochemical and cytogenetic profile.1, 2 FL3B is also a more aggressive subtype of FL and is known to behave similarly to high-grade lymphomas.3 Because FL3B is rare, data regarding its prognostication are lacking. It is unclear how positron emission tomography (PET)/computed tomography (CT) features might predict outcomes in the FL3B subtype. In a recent study, we found that extranodal (EN) and spleen involvement on PET/CT predict early clinical failure in FL grades 1–3A (FL1–3A).4 We aimed to determine whether this also applies to FL3B. PET/CT images from patients aged ≥18 years with newly diagnosed FL3B, between 2003 and 2016, were retrospectively reviewed (F.S.P. and S.M.B.). Patients were identified using the Mayo Clinic Lymphoma Database, as well as Mayo Clinic patients from the University of Iowa/Mayo Clinic Lymphoma SPORE Molecular Epidemiology Resource (MER) database. Patients were excluded if they had a previous diagnosis of lymphoma or inadequate pretreatment PET/CT imaging. Staging PET/CT examinations were uniformly reviewed by a trained medical resident (F.S.P.) and a nuclear medicine radiologist (S.M.B.). Bone marrow biopsies (BMB) were uniformly reviewed by one of our institutional haematopathologists (W.R.M.). Further details are available in Data S1. A total of 44 patients with FL3B were initially identified; 10 were excluded for inadequate pretreatment PET/CT images and seven were excluded because the FL3B was a recurrence of a previous lymphoma. In all, 27 patients were included in the study. All patients were treated with immunochemotherapy except two, who underwent surgical resection alone for localised disease. In all, 41% (11/27) of the patients had EN involvement on PET/CT; 30% (eight of 27) had one EN site, while 11% (three of 27) had two or more EN sites involved. The most common EN site involved was bone, in 30% (eight of 27) of the patients. The second most common was skin/soft tissue, in 15% (four of 27) of the patients. Brain, endometrium and liver involvement were each noted once. Evidence of spleen involvement on PET/CT was found in 41% (11/27) of the patients. In the present study cohort, 23/27 patients had BMB results available. Only four of the 23 patients had a positive BMB. We found that the sensitivity and specificity of bone involvement on PET/CT, compared to BMB, was 50% and 84% respectively. The sensitivity and specificity of spleen involvement on PET/CT, compared to BMB, was 50% and 68% respectively (Tables S1 and S2). The median (range) follow-up in patients still alive was 5·7 (0·1–15·6) years. In all, the mortality rate in our study cohort was 19% (five of 27) and an event occurred in 29·6% (eight of 27) of the patients. Biopsy-confirmed transformation occurred in 11% (three of 27) of the patients. The presence of bone involvement on PET/CT was associated with an increased risk of relapse (hazard ratio 6·01; P = 0·02). EN involvement and spleen involvement did not reach statistical significance for event-free survival (EFS) (Table I, Fig 1). Table I compares EFS results from this FL3B cohort to results obtained in the FL1–3A analysis.4 In our previous study on FL1–3A, 49% (301/613) of the patients had EN involvement on PET/CT, with 38% and 11% having one and two or more EN sites involved respectively.4 Spleen involvement was present in 28% (171/613) of the patients.4 Bone and skin/soft tissue were the two most common EN sites identified on PET/CT with involvement in 33% (204/613) and 10% (58/613) of the patients respectively.4 The sensitivity and specificity of PET/CT in determining bone involvement, compared to BMB, was 60% and 80% respectively, in FL1–3A.5 Spleen involvement on PET/CT also predicted bone marrow involvement on BMB, with a sensitivity and specificity of 55% and 86% respectively.5 Our present results indicate that the incidence of spleen and EN site involvement in patients with FL3B is similar to that of patients with FL1–3A. The distribution of EN sites was also similar. We found that the sensitivity and specificity of PET/CT in determining bone involvement on BMB in FL3B was similar to FL1–3A. Our findings indicate that PET/CT cannot replace BMB in determining marrow involvement, which differs from other lymphomas such as diffuse large B-cell lymphoma (DLBCL) or Hodgkin lymphoma. Among patients who underwent BMB, 17% (four of 23) had a positive BMB, which is similar to the usual rate of bone marrow involvement in DLBCL (c. 10%)6 and lower than that typically seen in FL1–3A (c. 35%).5 We found splenic involvement on PET/CT to be less strongly associated with BMB positivity as compared to our previous findings in FL1–3A. This may be due to the low number of positive BMB in our present cohort. In the present cohort of patients with FL3B, bone involvement on PET/CT was associated with inferior EFS (Fig 1). As such, factors that predicted early clinical failure in FL1–3A may also be predictive in FL3B, despite its distinct morphological and prognostic characteristics. In the FL1–3A study, the mortality rate, event rate and transformation rate were 11·3%, 38·2% and 8·8% respectively.4 We found that the event rate was higher in the FL1–3A cohort in comparison to FL3B. We hypothesise that this is secondary to the use of more aggressive immunochemotherapy at diagnosis, while various strategies were employed in patients with FL1–3A, including observation. The sample size may have also contributed to the low total event rate. Although aggressive initial treatment in FL3B appeared to result in lower rates of progression or transformation, patients with FL3B still had a poorer prognosis, with a mortality rate of 19% in our present FL3B cohort compared to 11% in FL1–3A. The strengths of our present study included a uniform review of imaging and pathology, consistency of treatment with immunochemotherapy across cases, and access to a large database of patients to identify our present cohort. Due to the rarity of FL3B, our present cohort size was small, which is a limitation to our study results. Only four patients had a positive BMB and the sensitivity/specificity data should be interpreted with caution. The number of events was insufficient to determine overall survival as a primary endpoint or to conduct a multivariate analysis on the subgroups identified as a higher risk of early clinical failure. A larger patient cohort, possibly achieved by pooled data from multiple centres, would be necessary to validate these findings. In the present cohort of 27 patients with FL3B, bone involvement on pretreatment PET/CT was associated with early clinical failure. Although FL3B behaves more aggressively than other FL grades, its prognostic factors on PET/CT at diagnosis may nonetheless be similar to that of lower grades. These findings may aid in the prognostication of patients with newly diagnosed FL3B, and in the guidance of their initial therapy. 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.