Making the Case for Maintenance Rituximab Owen A. O’Connor, M.D., Ph.D. owenoconnor@columbia.edu Professor of Medicine and Experimental Therapeutics Director, Center for Lymphoid Malignancies Columbia University Medical Center – College of Physicians and Surgeons The New York Presbyterian Hospital New York, N.Y.

Making the Case for Maintenance Rituximab The Problem The existing studies are almost as heterogeneous as the disease in question: - Are not consistent with regard to the kind of FL enrolled - Different doses and schedules of rituximab - Molecular heterogeneity ignored - Gender overlooked - Varying endpoints - Statistically never powered for OS

Heterogeneity of Follicular Lymphoma Morphology Guides Grading with Poor Concordance : Berard Criteria Grade 1 Grade 2 Grade 3 Small Cleaved Mixed Large Cell Large Cells per high powered field <5 5-15 >15% Expert concordance 72% 61% 60%

Heterogeneity of Follicular Lymphoma Follicular Lymphoma Classification Grade Characteristic Grade 1 0-5 centroblasts/hpf Grade 2 6-15 centroblasts/hpf Grade 3 > 15 centroblasts/hpf 3a Centrocytes present 3b Solid sheets of centroblasts Expert Pathologist Agreement With Consensus Diagnosis (ILSG, n = 304) Follicular Grade , % Dx 1 Dx 2 Dx 3 Any grade, 93 94 Grade 1 72 73 Grade 2 61 Grade 3 60 Swerdlow SH, et al. WHO classification of tumors of haematopoietic and lymphoid tissues (4th ed). Lyon, France: IARC Press; 2008.

Gene Expression: Follicular NHL Gene expression array demonstrates that the stromal microenvironment has profound prognostic influence Expression Signature Relative Risk of Death P Value Prognosis Immune response 1 0.15 < .0001 Favorable Immune response 2 9.35 Unfavorable Expression Signature (Prognosis) Relative Risk of Death P Value Immune response 1 (favorable) 0.15 < .0001 Immune response 2 (unfavorable) 9.35 Dave SS, et al. N Engl J Med. 2004;351:2159-2169.

Prognostic Factors in Follicular Lymphoma : Clinical Heterogeneity & FLIPI Age >60 Ann Arbor Stage III-IV LDH > Normal Hemoglobin < 12 Number of nodal sites >5 100 1 12 24 36 Months 48 60 72 84 20 40 80 Poor Intermediate Good Probability # of Factors Risk % of Patients 5 Year OS 10 Year OS 0-1 Good 36 91% 71% 2 Intermediate 37 78% 51% > 3 Poor 53% 36%

Monoclonal Antibodies Have Clearly Changed the Natural History of FL 100 CHOP + mAb 80 Pro-MACE 60 Patients (%) 40 N Deaths, n 4-Yr Estimated OS, % CHOP 20 179 18 91 425 189 79 356 226 69 2 4 6 8 10 Yr After Registration Fisher RI, et al. J Clin Oncol. 2005;23:8447-8452.

Rationale for Maintenance Therapy in Indolent Lymphoma Maintenance therapy applied in patients responding to induction treatment is effective in hematological malignancies Maintenance therapy can deepen the response and lengthen remission Need to have therapeutic agents with a good efficacy/toxicity ratio: No cumulative toxicity (hematopoietic stem cells) No long term side effects Preserve quality of life Do no compromise subsequent treatment(s) efficacy

The Variability of Maintenance Rituximab Studies: Across Many Designs Same Conclusion Study Population Treatment Findings SAKK 35/98 Blood 2004; 103(12) JCO 2010; 28(29) Heterogeneous. Prior chemo (n=138) and chemonaive (n=64); all grades; MAB naive Rituximab 375 mg/m2 per week x 4 then randomize to observation vs 4 more doses (months 3, 5, 7 and 9) Median F/U of 9.5 yrs EFS 13 vs 24 months 45% of untreated patients EF @ 8 yrs stage not prognostic; Fc-gamma RIIIA Hainsworth JCO 2005; 6(20) FL Gr 1 or 2 and SLL (23 and 39% of patients); 1 prior chemotherapy Rituximab 375 mg/m2 per week x 4 then randomize to retreatment on POD or maintenance weekly x 4 Q 6 months till POD or total of 4. PFS was 31.3 vs 7.4 months (ORR and CR >> in maintenance arm) EORTC 20981 (Van Oers) JCO 2010; 28(17) FL Gr 1-3; Stage 3-4 only; relapsed after max of 2 non-anthracycline regimens. Randomized to CHOP vs R-CHOP then randomized maintenance rituximab 375 mg/m2 once Q 3 months Rituximab maintenance PFS 3.7 vs 1.3 years. 5 yr OS 74% vs 64% (p=0.07). Increase in infections (9.7 vs 2.4%) German Low Grade Lymphoma Study Group Forstpointner Blood 2006; 108(13) FL all grades; relapsed after 1 prior; FCM vs R-FCM; then randomize to maintenance weekly x 4 at 3 and 9 months or observation PFS and response duration not met in maintenance arm vs 26 months in observation. (median observation of 26 months) PRIMA (Salles) Lancet 2011; 377 Untreated FL, Grades 1-3a; high tumor burden (one lesion > 7 cm; 3 nodes > 3 cm; symptomatic splenomegaly; organ compression; effusions; increased LDH or b2 or ‘B’ symptoms) R-Chemo (CVP/CHOP/FCM) followed by rituximab 375 mg/m2 every 8 weeks for two years. PFS at 6 years 59% vs 42% favors maintenance, suggestion less transformation, no difference is OS. Meta-Analysis All studies Variable Highly statistically significant improvement in HR (0.005) for overall survival.

Arm B Rituximab induction Arm C Rituximab induction and maintenance Preliminary Analysis of Rituximab vs Watch and Wait in Asymptomatic FL Patients R A N D O M I Z A T N Arm A Watch and wait Clinic visits R x 4 Patients with FL grades 1, 2, and 3a, stage II, III, IV disease, ECOG PS 0-1 (N = 462) Arm B Rituximab induction Continued follow-up R x 4 R R R R R R R R R R R R Arm C Rituximab induction and maintenance 1 3 5 7 9 11 13 15 17 19 21 23 25 Mos Compulsory CT Scan CT Scan only if clinical CR Compulsory CT Scan PD requiring therapy stops protocol treatment Bone marrow for histology and MRD only if CT shows CR Ardeshna K, et al. ASH 2010. Abstract 6.

Preliminary Analysis of Rituximab vs Watch and Wait in Asymptomatic FL Patients Spontaneous remission observed in 3% of patients on watch and wait vs CR in 45% of patients on rituximab 93 patients required new therapy during follow-up period 84 patients (90%) had PD 78 patients (84%) received chemotherapy as new treatment 3-Yr PFS W + W: 33% R4: 60% R4 + RM: 81% Progression Free Survival 1.0 0.8 Proportion of Patients Progression Free 0.6 0.4 Events 108 33 Totals 181 83 189 0.2 W + W R4 R4 + RM 1 2 3 4 5 Yrs From Randomization HR (rituximab vs W + W): 0.46; 95% CI: 0.33-0.65; P < .001 HR (rituximab + M vs W + W): 0.21; 95% CI: 0.15-0.29; P < .001 HR (rituximab + M vs rituximab): 0.43; 95% CI: 0.24-0.72; P = .001 Ardeshna K, et al. ASH 2010. Abstract 6.

Rituximab Maintenance Therapy in FL The First Study (SAKK Trial) Observation (n=78) R A N D O M I Z E Rituximab 375 mg/m2 qw × 4 CR, PR, SD maintenance 375 mg/m2 q8w  4 (n=73) Untreated or relapsed/ refractory FL grade I-III (N=202) ASSESS Ghielmini et al. Blood. 2004;103:4416.

Rituximab Maintenance Therapy in FL (SAKK Trial): Event Free Survival Time from 1st induction to POD, relapse, second CA or death 1.0 0.8 Prolonged: median 23.2 months 0.6 Probability 0.4 0.2 Standard: median 11.8 months P=0.024 6 12 18 24 30 36 42 48 Months Ghielmini et al. Blood. 2004;103:4416.

Updated EFS in SAKK 35/98: Rituximab Maintenance vs. Observation

Overall Survival: ASCO update of SAKK 2009 Probability P = 0.09 Is this study powered for OS endpoint? NO! Years Since Start of Treatment

Rituximab 375 mg/m2 weekly × 4 Rituximab Maintenance Therapy vs Re-Treatment at Progression for Indolent NHL 114 rituximab-naive patients with previously treated indolent NHL Rituximab 375 mg/m2 weekly × 4 Rituximab maintenance therapy (n=44) Rituximab 375 mg/m2 weekly × 4 every 6 months for 4 courses R A N D O M I Z E 90 patients (70%) CR/PR/SD Re-treatment at time of progression (n=46) Rituximab 375 mg/m2 weekly × 4 Hainsworth et al. Blood. 2003;102(11). Abstract 231

LYM-5 - Maintenance vs Retreatment After Rituximab : Hainsworth Regimen FL, SL rel/refr; R weekly x 4 + (repeat every 6 months x 4 max) or (treat at progression) Progression Free Survival 20 40 60 80 100 24 Months 36 48 P=0.007 Maintenance Retreatment 12 % Surviving w/o progression Hainsworth et al. J Clin Oncol. 2005;23:1088

Phase III Trial of Rituximab Maintenance for Patients in Remission with Relapsed / Resistant Follicular NHL R A N D O M I Z E R A N D O M I Z E CHOP every 21 days maximum 6 cycles Observation CR PR Rituximab + CHOP every 21 days maximum 6 cycles Rituximab maintenance* *375mg/m2 q 3 months x 2 yrs or until relapse Van Oers et al. Blood. 2006;108:3295

Median PFS after CHOP (mo) Median PFS after R-CHOP (mo) R-Maintenance in Patients With Relapsed/ Resistant FL After CHOP or R-CHOP: Progression Free Survival Median PFS after CHOP (mo) R-maintenance 42.2 Observation 11.6 Median PFS after R-CHOP (mo) P=0.004 HR 0.54 P<0.001 HR 0.30 R-Maintenance 51.8 Observation 23.0 10 20 30 40 50 60 70 80 90 100 1 2 3 4 5 Years Percent Van Oers et al. Blood. 2006;108:3295.

Update of EORTC 20981 Intergroup Phase III Trial in R/R Follicular Lymphoma Patients: PFS from 2nd randomization van Oers M, et al. J Clin Oncol. 2010;28(17):2853-8.

EORTC 20981 Intergroup phase III trial: 5-year Overall Survival Again, trends towards significance, despite being underpowered for this endpoint Is it too early for this disease? van Oers M, et al. J Clin Oncol. 2010;28(17):2853-8.

PRIMA Study Design : High Tumor Burden INDUCTION MAINTENANCE Rituximab maintenance 375 mg/m2 every 8 weeks for 2 years Registration High tumor burden untreated follicular lymphoma Immunochemotherapy 8 x rituximab + 8 x CVP or 6 x CHOP or 6 x FCM CR/CRu PR Random 1:1 Observation PD/SD off study 5 YEARS FOLLOW-UP Salles G, et al. Lancet. 2011;377(9759):42-51.

Primary Endpoint (PFS) Met at the Planned Interim Analysis Rituximab maintenance significantly reduced the risk of lymphoma progression by 50% (stratified by response and induction regimen, HR=0.50, 95% CI 0.39; 0.64) 1.0 82% Rituximab maintenance N=505 0.8 0.6 Observation N=513 HR=0.50 p<0.0001 66% Progression-free rate 0.4 0.2 6 12 18 24 30 36 42 Patients at risk Time (months) 505 472 443 336 230 103 18 513 469 411 289 195 82 15 Salles G, et al. Lancet. 2011;377(9759):42-51.

Subgroup Analyses Results Hazard ratio * Category Subgroup Hazard ratio N 95% CIs All All 1018 0.49 0.38 0.64 Age < 60 624 394 0.45 0.59 0.33 0.39 0.62 0.90 ≥ 60 Sex Female 485 533 485 533 0.58 0.43 0.40 0.31 0.85 0.61 Male FLIPl Index (CRF) FLIPl ≤ 1 216 370 431 0.38 0.39 0.61 0.19 0.25 0.43 0.77 0.61 0.87 FLIPl = 2 FLIPl ≥ 3 R-CHOP Induction Chemotherapy 768 222 28 0.43 0.69 0.51 0.31 0.44 0.13 0.59 1.08 2.07 R-CVP R-FCM Response to Induction CR/CRu 721 290 0.52 0.45 0.38 0.29 0.70 0.72 PR 1 2 3 Favors maintenance Favors observation * Non-stratified analysis Salles G, et al. Lancet. 2011;377(9759):42-51.

Efficacy Across Secondary Endpoints Substantial Improvement with R-Maintenance Time to next anti-lymphoma treatment Time to next chemotherapy treatment Rituximab maintenance Rituximab maintenance 1.0 1.0 0.8 0.8 0.6 0.6 Event-free rate Event-free rate HR = 0.61 p = 0.0003 Observation HR = 0.60 p = 0.0011 Observation 0.4 0.4 0.2 0.2 6 12 18 24 30 36 42 6 12 18 24 30 36 42 Time (months) Time (months) Patients at risk 513 487 447 327 218 87 15 513 492 454 332 225 91 17 505 483 453 349 235 103 18 505 484 457 351 243 108 19 Salles G, et al. Lancet. 2011;377(9759):42-51.

PRIMA 6-Years Follow-Up : PFS From Randomization PFS according to maintenance (ITT patients) With number of subjects at risk and 95% confidence intervals 1.0 – 0.8 – 0.6 – 0.4 – 0.2 – 0.0 – Observation Rituximab Survival Probability 6 years = 59.2% HR = 0.57 P<.0001 6 years = 42.7% 0 10 20 30 40 50 60 70 80 90 PFS Delay Observation Rituximab 513 438 361 302 273 240 210 140 36 0 505 456 418 387 351 328 298 188 50 0 No of Subjects Event Censored Median Survival (95% CI) Observation 513 56.5% (290) 43.5% (223) 8.5 (41.2 - 59.4) Rituximab 505 39% (197) 61% (308) NA (82.6 - NA) Median follow-up since randomization: 73 months Salles G, et al. Blood. 2013;122: Abstract 509.

PRIMA 6-Years Follow-Up : Overall Survival OS according to maintenance (ITT patients) With number of subjects at risk and 95% confidence intervals 1.0 – 0.8 – 0.6 – 0.4 – 0.2 – 0.0 – 6 years = 88.7% Survival Probability 6 years = 87.4% HR = 1.027 P = .885 Observation Rituximab 0 10 20 30 40 50 60 70 80 90 OS Delay Observation Rituximab 513 503 491 478 468 456 431 291 72 0 505 492 482 471 458 447 423 288 80 0 No of Subjects Event Censored Median Survival (95% CI) Observation 513 11.3% (58) 88.7% (455) NA (NA - NA) Rituximab 505 11.7% (59) 88.3% (446) Median follow-up since randomization: 73 months Salles G, et al. Blood. 2013;122: Abstract 509.

Overall Survival: Meta-analysis R maintenance in FL Patients In First Line: Minimal effect on overall survival OR Shorter follow-up and limited power of the studies (low number of deaths ?) Vidal et al. J Natl Cancer Inst. 2011 Dec 7;103(23):1799-806

ECOG 4402 (RESORT) Primary End-point - Time to treatment failure Accruing 389 patients with low-tumor-burden stage III/IV indolent NHL Rituximab maintenance 375 mg/m2 q12w R A N D O M I Z E Rituximab 375 mg/m2 qw  4 CR or PR Rituximab re-treatment at progression 375 mg/m2 qw  4 Primary End-point - Time to treatment failure Secondary endpoint - Time to first cytotoxic therapy http://www.clinicaltrial.gov/ct/show/NCT00075946?order=2. Accessed May, 2005.

Is The Resort Study an Appropriate Comparator? Population Treatment Findings SAKK 35/98 Blood 2004; 103(12) JCO 2010; 28(29) Heterogeneous. Prior chemo (n=138) and chemonaive (n=64); all grades; MAB naive Rituximab 375 mg/m2 per week x 4 then randomize to observation vs 4 more doses (months 3, 5, 7 and 9) Median F/U of 9.5 yrs EFS 13 vs 24 months 45% of untreated patients EF @ 8 yrs stage not prognostic; Fc-gamma RIIIA Hainsworth JCO 2005; 6(20) FL Gr 1 or 2 and SLL (23 and 39% of patients); 1 prior chemotherapy Rituximab 375 mg/m2 per week x 4 then randomize to retreatment on POD or maintenance weekly x 4 Q 6 months till POD or total of 4. PFS was 31.3 vs 7.4 months (ORR and CR >> in maintenance arm) EORTC 20981 (Van Oers) JCO 2010; 28(17) FL Gr 1-3; Stage 3-4 only; relapsed after max of 2 non-anthracycline regimens. Randomized to CHOP vs R-CHOP then randomized maintenance rituximab 375 mg/m2 once Q 3 months Rituximab maintenance PFS 3.7 vs 1.3 years. 5 yr OS 74% vs 64% (p=0.07). Increase in infections (9.7 vs 2.4%) German Low Grade Lymphoma Study Group Forstpointner Blood 2006; 108(13) FL all grades; relapsed after 1 prior; FCM vs R-FCM; then randomize to maintenance weekly x 4 at 3 and 9 months or observation PFS and response duration not met in maintenance arm vs 26 months in observation. (median observation of 26 months) PRIMA (Salles) Lancet 2011; 377 Untreated FL, Grades 1-3a; high tumor burden (one lesion > 7 cm; 3 nodes > 3 cm; symptomatic splenomegaly; organ compression; effusions; increased LDH or b2 or ‘B’ symptoms) R-Chemo (CVP/CHOP/FCM) followed by rituximab 375 mg/m2 every 8 weeks for two years. PFS at 6 years 59% vs 42% favors maintenance, suggestion less transformation, no difference is OS. Meta-Analysis All studies Variable Highly statistically significant improvement in HR (0.005) for overall survival. RESORT FL Gr 1-2 (+SLL, MZL); low tumor burden; Stage III-IV, no prior treatment Rituximab 375 mg/m2 weekly x 4 then randomized to same with progression vs one dose Q 13 weeks. TTTF equivalent in RR vs MR though time to next cytotoxic treatment favors MR

The Importance of Study Endpoints Time to Treatment Failure Time to treatment failure (TTF) is defined as the time from randomization to treatment discontinuation for any reason, including disease progression, treatment toxicity, patient preference, or death. From a regulatory point of view, TTF is generally not accepted as a valid endpoint. TTF is a composite endpoint influenced by factors unrelated to efficacy. Discontinuation may be a result of toxicity, patient preference, or a physician's reluctance to continue therapy. These factors are not a direct assessment of the effectiveness of a drug

The Importance of Study Endpoints Progression Free Survival The progression-free survival (PFS) duration is defined as the time from randomization to objective tumor progression or death. Compared with other endpoints, PFS is a preferred regulatory endpoint because it includes death and may correlate better with OS. Assessment of either PFS or TTP needs to be conducted in randomized trials. To reduce bias, the same assessment technique should be used at each follow-up, and the same evaluation schedule should be consistently used.

Maintenance Rituximab (R): To Give or Not To Give? Several different schedules of R maintenance have been or are being tested Optimal dose/schedule and dosing is still unknown! Men receiving RM do worse than women… PK study? How important is it to determine the smallest amount of R & optimal time-points that is associated with improved PFS?

Maintenance Rituximab in Follicular Lymphoma : Where Do We Stand Follicular lymphoma is a remarkably heterogeneous disease, and no study has addressed the value of MR as a function of this heterogeneity There is enormous variability across all studies with regard to the study population, at least with respect to treatment, grade, stage, and little data on FLIPI, none on GEP Every study consistently demonstrates compelling prolongation of PFS, some trending close to OS (not mature enough for the end-point) No study ever powered for OS! What’s the best end-point? Is TTTF a meaningful end-point in FL studies? I don’t think so, profoundly confounded.

Maintenance Rituximab in Follicular Lymphoma : Where Do We Stand Gender differences are now emerging in clinical studies of rituximab, demonstrating substantial variability in PK between male and female. Would adjustment of dose based on gender make a difference? PFS is a clinically meaningful end-point (per FDA) in FL. Who wants to move to cytotoxic therapy sooner? RESORT demonstrates MR favors longer time to cytotoxic therapy Building time between cytotoxic therapies may be an important determinant of chemotherapy sensitivity to future lines of treatment Personally, I tailor the recommendation for maintenance rituximab based to the patient circumstances and their sensitivity to previous rituximab

Thank You

PRIMA: Statistical Assumptions Sample size calculation: Based on a 45% increase in median PFS, with a power of 80% to detect this difference (Type I alpha risk - 2-sided - of 5%) 1200 patients registered at induction, estimated response rate of 75%: 900 randomized to maintenance or observation (1:1) Full analysis after 344 events Interim analysis planned after 258 events: Stopping rules for the IA: two-sided O’Brien-Fleming boundary of 2.3397 with type I of 5% (nominal p = 0.0193) Submitted to an independent Data Safety Monitoring Board

Maintenance Therapy for Indolent and Mantle Cell Lymphoma Unanswered Questions PK data support the scheme with one administration every 2 months Benefit of longer duration of maintenance remains to be determined

PRIMA: Patient Disposition Induction Patients registered: N = 1217 * 3 sites closed prematurely (15 pts) Patients evaluable (N = 1202) * . 9 pts did not received chemo . 147 pts withdraw during or at the end of induction (failure to respond; toxicity) . 28 pts failed to be randomized R-CHOP N = 885 R-CVP N = 272 R-FCM N = 45 Randomized N = 769 Randomized N = 222 Randomized N = 28 Maintenance Patients randomized: N = 1018 * * 1 patient died during the randomization process Observation N = 513 Rituximab N = 505 Salles G, et al. Lancet. 2011;377(9759):42-51.