Risk and Prognostic Factors of Extra-Nodal Hodgkin Lymphoma: Single Jordanian Center Experience

Background

Hodgkin lymphoma (HL) is a clonal B-cell hematologic malignancy usually affecting only the lymphoid tissue, as extra-nodal (EN) involvement (e.g., bone, lung or liver) is seen only in ~15% of cases, much less than what is observed in non-Hodgkin lymphomas. To date, a small number of studies worldwide have investigated the epidemiology and clinical characteristics associated with HL with EN manifestations.

Aim

To describe risk factors and prognosis of EN involvement in HL patients.

Methods

Patients diagnosed with HL at King Abdullah University hospital (KAUH) between January 2004 and April 2018 were retrospectively reviewed for EN involvement, confirmed by histopathology of involved tissues. Risk factors assessed in our cohort of EN HL patients were: age, gender, presenting symptoms, B Symptoms, LDH levels, histological subtypes, CBC with differential, CT scans and PET/CT scans. International prognostic score (IPS) was also calculated.

Results

Among 140 HL patients included in our study, 16 of them (11.4%) had EN involvement. The mean age at diagnosis was 30 years old, ranging from 14-59. As in classic HL, men with EN involvement were slightly more affected than female (M/F, 10/6; 62.5% vs. 37.5%). Most of patients (N=8; 50%) had HL with mixed cellularity (MC); 6 patients (37.5%) had nodular sclerosing (NS); and 2 patients (12.5%) were classified as lymphocyte depleted (LD) subtype. Bone was the most compromised extra-nodal tissue as shown by the observation that 8 patients (50%) had bone involvement. Lungs pleura were affected in 7 patients (43.8%) and liver in 4 (25%) patients. The remaining subjects (N=4; 25%) showed involvement in other sites (skin, CNS and muscles). Out of our 16 EN HL patients, 6 of them (37.5%) had two EN tissues involved.

EN HL patients were younger than patients without EN involvement (mean age, 30 vs. 37 years old respectively, P=.036 by unpaired t-test).EN HL patients with bone involvement were younger than those with liver or lung localization (mean age, 23.1 vs. 30.8 vs. 30.6 years old, respectively; P = 0.018 by unpaired t-test).

We also observed a significant association between liver involvement and MC subtype. (P =0.032 by chi-square test).

There was significant association between positive B symptoms and EN involvement. (P=.006 by chi-square test). There was no significant association between gender, high LDH levels, anemia, leukocytosis, lymphocytosis, monocytopenia, low albumin levels and EN involvement.

There was significant association between IPS and EN involvement (56.25% of patients with EN involvement have high risk IPS, P=.00002 by chi-square test).

All patients received ABVD as a first-line treatment. Seven patients (43.75%) experienced disease relapse during the follow-up period. There was a significant relationship between disease relapse and the type of HL, as 6 patients who experienced relapse had NS (P=0.0004 by chi-square), while only one MC.

Overall survival rate was 75%, and 5-years survival rate was 45%. Four out of 7 patients who had disease relapse (57.1%) died, suggesting that disease relapse in EN HL might be a poor prognostic factor (P =0.045 by Kaplan Meier). There was no significant association between EN involvement and overall survival rate (P=.35 by Kaplan Meier).

Conclusion

Prevalence of EN involvement in patient with HL in Jordan is 11.4%, higher than that in Western countries and United States. In our cohort, EN involvement frequently presented as MC histological subtype; however, patients with NS showed the highest relapse rate. According to previous observations, bone and lungs were the most involved sites in our cohort, and patients with bone involvement were younger than those with other extra-nodal sites. Of note, our study might suggest some differences in the primary disease site in patients with HL possibly due to some geographic and ethnic variations. We also showed that risk and prognostic factors of classic HL are well applicable to EN HL, highlighting possible roles of histological subtypes in prognostication of EN HL. However, our preliminary findings need further validation in larger prospective studies.