Non-Hodgkin lymphoma impacts a patient's physiology beyond manifesting as lymphadenopathy, which was recognized in the original Ann Arbor staging such as B-symptoms. Likewise, cancer cachexia is prevalent across tumor types, but it is felt to be a late feature of advanced disease. We have previously demonstrated that cancer patients may have muscle changes well before cachexia's onset such as a reduced muscle radiation attenuation as seen on computed tomography (CT) imaging. This drop in skeletal muscle radiographic density (SMD) was shown to be prognostic and possibly predictive of rituximab-containing treatment response in retrospective studies of follicular and diffuse large B-cell lymphomas1. This post hoc analysis of LYM 3002 (NCT00722137), a phase III randomized trial of newly diagnosed mantle cell lymphoma (MCL) patients, sought to investigate SMD's impact in both MCL and validate it in prospectively collected data.


487 newly diagnosed MCL patients were randomized 1:1 between R-CHOP or VR-CAP (vincristine or bortezomib at 1.3 mg/m2 d1, 4, 8, 11 q3wk) for a total of 6-8 cycles in LYM 30022. CT images at pre-treatment, 12-weeks post treatment, and 26-weeks post treatment were anonymized (for blinding) and reviewed to measure patient SMD at the L3 vertebral body level. Based on prior retrospective data, normal SMD was defined as ≥ 41 and 33 Hounsfield Units for non-overweight (BMI < 25 kg/m2) and overweight (BMI ≥ 25 kg/m2) patients, respectively. Patients with normal vs. low SMD were compared for differences in investigator-assessed progression free (PFS) and overall survival (OS). Cox proportional hazards multivariate analysis weighed SMD's significance against other known prognostic factors: gender, MIPIb score (accounting for Ki-67 score), and treatment received.


Of 487 patients, 478 had CT images available for review and were split equally between either treatment arm. 42.5% (n=203) were found to have low SMD at baseline. At a median follow-up of 40 months, patients with low vs. normal baseline SMD had substantially poorer, investigator-assessed, median PFS (16.3 vs. 23.7 months, hazard ratio [HR] 1.52, p<0.001) in both treatment groups combined. Though not yet mature for comparison between treatment groups, OS was similarly impacted between SMD groups (HR 1.85, p<0.001) with 3-year OS rate was 61.6 vs. 76.8%, respectively. The presence of low vs. normal SMD at 12 or 26 weeks post treatment completion was still prognostic of PFS (n=371, HR 1.54, p=0.001 and n=317, HR 1.43, p=0.02, respectively) and OS (HR 1.89, p=0.004 and 1.85, p=0.02, respectively). In multivariate analysis taking into consideration gender, MIPIb score and treatment received, low baseline SMD retained its negative impact on OS (HR 1.86, p=0.005) and trended to significance on PFS (HR 1.27, p=0.12) owing partially to improved PFS seen in patients receiving VR-CAP. SMD status at 12 and 26 weeks post treatment displayed similar prognostic impact. Though low SMD patients receiving VR-CAP had improved median PFS compared to those receiving R-CHOP (22.4 vs. 14.4 months, HR 0.66, p=0.02), the presence of low vs. normal SMD still impacted outcomes of VR-CAP patients (22.4 vs. 36.2 months, HR 1.64, p=0.007). From baseline to 12 weeks post treatment in both arms, 27 patients with low SMD were able to regain muscle density to normal levels while 68 moved from normal to low SMD. A comparison of median PFS based on change in SMD over time showed a stepwise trend from those who remained with low SMD throughout, those who developed low SMD, those who became normal, and those who stayed with normal SMD throughout (23.7 vs. 25.4 vs. 29.4 vs. 37.5 months, respectively).


This post hoc analysis demonstrates and validates SMD's prognostic ability in MCL regardless of therapy similar to retrospective findings in follicular and diffuse large B-cell lymphoma. We also demonstrate that not only can SMD change post treatment, but this change may predict for better outcomes. By using routine CT images, SMD assessment can be a strong, inexpensive, and simple complement to MIPIb prognostication. Low SMD is hypothesized to represent a specific inflammatory cytokine milieu such as seen in cancer cachexia. Whether this same inflammatory phenomenon represents the SMD changes seen in lymphoma is currently being investigated.

1 PLOS One 2015;doi:10.1371/journal.pone.0127589

2 N Engl J Med 2015;372(10):944-53.


Belch:Janssen-Cilag: Consultancy. Pei:Janssen Research and Development: Employment; Johnson & Johnson: Other: Stock Ownership. Yang:Janssen Research and Development: Employment. Robak:Janssen: Consultancy, Research Funding; MorphoSys AG: Consultancy, Honoraria, Research Funding. Rooney:Janssen Research and Development: Employment; Johnson & Johnson: Other: Stock Ownership. Cakana:Janssen Research and Development: Employment. van de Velde:Millennium Pharmaceuticals, Inc.: Employment; Johnson & Johnson: Equity Ownership.

Author notes


Asterisk with author names denotes non-ASH members.