Thymectomy in Non-thymomatous MG: Results from MTGX, a Randomized, Controlled Trial MSG Scientific Annual Meeting Sept 2016, Snowbird, UT Gil I. Wolfe, MD Dept. of Neurology Univ. at Buffalo/SUNY Jacobs School of Medicine and Biomedical Sciences Buffalo, NY U01 NS042685

The Thymectomy Dilemma 1) Why? 2) How? (MGTX trial) 3) What? (MGTX trial results)

The Thymectomy Dilemma 1) Why?

Thymectomy for non-thymomatous MG Historical aspects Blalock et al. JAMA 1941;117:1529-1533 Sustained remission in initial patient with cystic thymic tumor (Blalock et al. Ann Surg 1939;110:544-561) Removal of entire thymus gland from mediastinum in 6 patients without thymoma 1 pt symptom-free 2 pts significantly improved 2 pts slightly improved 1 death from pulmonary complications No Class I evidence for over 70 years

Thymectomy for MG Non-thymomatous MG Consensus in favor of thymectomy Based on non-randomized case series Repeated calls for randomized, controlled studies Gronseth & Barohn, 2000 (Neurology 55:7) Grob, 1981 (comments in Ann NY Acad Sci 377:764) McQuillen & Leone, 1977 (Neurology 27;1103) Trial efforts failed in 1963, 1980, early 1990s

Lifetime Course of MG Grob et al Lifetime Course of MG Grob et al. Muscle Nerve 2008;37:141 David Grob, MD, 1919-2008 Transsternal thymectomy effect on remission 1940-57 Significant effect (20% vs. 10%) 1958-1967 Similar remission and improvement rates 1966-2000 Slightly higher mortality and lower remission rates in thymectomy group P values vs. 1940-57

Non-thymomatous MG Evidence-based AAN Practice Parameter Gronseth & Barohn. Neurology 2000;55:7 28 Class II studies (controlled, non-randomized) 21 MG cohorts 4136 surgical pts 4354 non-surgical pts Published between 1953-1998 Majority used transsternal approach Mean f/u range 3-28 years No blinded assessments

Non-thymomatous MG AAN Practice Parameter Gronseth & Barohn. Neurology 2000;55:7

AAN Practice Parameter Conclusions and recommendations Benefit of thymectomy not established conclusively More effective in generalized, severe, female patients? Recommended as treatment option Controlled trial needed Standardized medical therapy for all arms Defined outcome measures Compare different surgical approaches Gronseth & Barohn. Neurology 2000;55:7

The Thymectomy Dilemma 1) Why? 2) How?

MGTX and BioMG Executive Committee U01 NS042685 John Newsom-Davis, MD (Study Chair, emeritus) Gary Cutter, PhD (Director, DCC, UAB) Gil Wolfe, MD (Study Chair) Henry Kaminski, MD (Study Vice Chair; Director BioMG) Inmaculada Aban, PhD (Deputy Director, DCC) Alfred Jaretzki, MD (Surgical Chair, emeritus) Joshua Sonnett, MD (Surgical Chair) Greg Minisman (Project Manager) Robin Conwit, MD (NINDS Program Director) Joanne Odenkirchen, MPH (NINDS colleague) 1932-2007 U01 NS042685 1919-2014

MGTX: Single blind, Multicenter, International Randomized Trial Primary aim: Answer 3 questions in non-thymomatous MG patients followed for 3 years Compared with prednisone protocol alone, does extended transsternal thymectomy (ETTX) + prednisone protocol result in: greater improvement in strength (time-weighted average QMG)? lower prednisone requirement (time-weighted average prednisone dose)? enhanced quality of life (AEs, TAS, TAC)?

MGTX protocol details Inclusion criteria Main exclusion criteria AChR binding Ab pos (≥0.5) MGFA Class 2-4; disease duration < 5 years Age at least 18 and < 65 years Optimal anti-cholinesterase dose Prednisone naive or not Main exclusion criteria Previous thymectomy or sternotomy or thoracotomy Immunosuppressive therapy (x prednisone) within last year Rituximab at any time Medically or psychiatrically unfit for thymectomy Chest CT or MR evidence of thymoma Pregnancy or lactation, or considering becoming pregnant Current prednisone > 0.75 mg/kg or 50 mg/d (or AD equivalent)

MGTX Trial 67 sites in N. America, Europe, S. America, S. Africa, Asia, Australia AChRAb+, MGFA Clinical Class II-IV, ≤5 yrs, no thymoma, 18-65 yo, +/- prednisone Rx ETTX + Prednisone 1.5 mg/kg AD randomize Prednisone alone 1.5 mg/kg AD MMS: prednisone taper MMS: prednisone taper 1° Time-weighted QMG & Prednisone, AEs at 3 yrs 2° MMS MG-ADL ∆SF-36 Hospital days 1 ° Time-weighted QMG & Prednisone, AEs at 3 yrs 2° MMS ∆MG-ADL ∆SF-36 Hospital days outcome measures

MGTX protocol details If primary outcome ambiguous Dual primary outcome Clinical response (Time-weighted QMG) Blinded evaluator (BE) Prednisone requirements (Time-weighted AD prednisone) If primary outcome ambiguous Serious adverse events/TAC/TAS Prednisone dosing based on Minimal Manifestation Status (MGFA) No symptoms or functional limitations from MG but some weakness present on careful examination Jaretzki et al. Neurology 2000;55:16 QMG score <14; ≤ baseline value Treatment options Azathioprine or other IS agent allowed if MMS not reached by 1 year OR severe prednisone-related AEs via protocol deviation

Investigator Predictions Prior to Starting the Trial

MGTX Results 1) Why? 2) How? 3) What? Wolfe GI et al. N Engl J Med 2016;375:511-522

Rate of Recruitment

MGTX: n=126 Refusals ITT analysis

Thymectomy+prednisone Demographic and Baseline Clinical Characteristics of MGTX Subjects   Characteristic Prednisone Alone (N=60) Thymectomy+prednisone (N=66) Gender – no. (%) Female 39 (65) 50 (76) Ethnicity – no.(%)  Asian 4 (7) 6 (9) Black/African American 6 (10) 7 (11) Hispanic 17 (28) 17 (26) White, not Hispanic origin 30 (50) 31 (47) Other (Mixed/Native American/Alaskan) 3 (5) 5 (8) Therapy at enrollment* – no (%) Pyridostigmine 56 (93) 60 (91) Corticosteroids 47 (78) 49 (74) Prior intravenous immunoglobulin 13 (22) 12 (18) Prior plasma exchange 7 (12) 9 (14) MG Foundation of America Class† – no.(%) Class IIa 25 (42) 25 (38) Class IIb 14 (23) 18 (27) Class III 20 (33) 21 (32) Class IV 1 (2) 2 (3) Median age in years (min-max) 33.0 (18.0-64.0) 32.0 (18.0-63.0) Median disease duration in years (min-max) 1.14 (0.15-4.38) 1.08 (0.02-4.41) Quantitative MG score‡: Mean ± SD 12.35 ± 4.90 11.40 ± 5.12 Prednisone dosage (mg): Mean ± SD 42.49 ± 23.52 43.43 ± 28.92

QMG Score (Mean±SE) by Treatment Group QMG difference: 2.85 pts (99.5% CI 0.47-5.22; p<0.001)

AD Prednisone Dose (Mean±SE) by Treatment Group Time weighted average prednisone dose difference: 44 mg vs 60 mg (95% CI 7-25 mg; p<0.001)

Primary and Subgroup Analyses   Treatment Group Mean±SD Estimated Difference (95% CI†) P Valuea Prednisone alone Thymectomy+ prednisone Primary Analyses Time-weighted average Quantitative MG score 8.99 ± 4.93 (N=56) 6.15 ± 4.09 (N=62) 2.85 (0.47-5.22) < 0.001 alternate-day prednisone dose (mg) 59.8 ± 27.4 (N=56) 43.6 ± 21.2 (N=61) 16.2 (7.3-25.1) Subgroup Analyses Time-weighted average Quantitative MG score Prednisone use at enrollment (interaction with treatment P valueb=0.86) Not prednisone naїve 9.10 ± 5.06 (N=46) 6.30 ± 3.89 (N=47) 2.80 (0.11-5.49) 0.004 Prednisone naїve 8.84 ± 4.60 (N=9) 5.66 ± 4.79 (N=15) 3.18 (-3.03-9.39) 0.12 Gender (interaction with treatment P valueb=0.57) Female 9.73 ± 5.16 (N=38) 6.47 ± 4.13 (N=46) 3.26 (0.34-6.18) 0.002 Male 7.45 ± 4.11 (N=18) 5.23 ± 3.95 (N=16) 2.22 (-1.96-6.40) Age (years) at disease onset (interaction with treatment P valueb=0.74) < 40 9.60 ± 5.32 (N=34) 6.50 ± 4.41 (N=42) 3.10 (-0.13-6.33) 0.007 ≥ 40 7.85 ± 3.50 (N=18) 5.33 ± 2.79 (N=18) 2.52 (-0.65-5.69) 0.02 Time-weighted average alternate-day prednisone dose (mg) Prednisone use at enrollment (interaction with treatment P valueb=0.37) 61.5 ± 28.5 (N=46) 44.3 ± 21.9 (N=46) 17.2 (6.7-27.7) 48.5 ± 19.0 (N=9) 41.5 ± 19.6 (N=15) 7.0 (-10.0-24.0) 0.40 Gender (interaction with treatment P valueb =0.32) 59.19 ± 24.86 (N=37) 45.76 ± 22.80 (N=45) 13.43 (2.95-23.92) 0.01 61.08 ± 32.46 (N=19) 37.70 ± 14.94 (N=16) 23.38 (6.26-40.50) 0.009 Age (years) at disease onset (interaction with treatment P valueb=0.94) 60.79 ± 27.11(N=33) 44.92 ± 21.95 (N=41) 15.87 (4.50-27.23) 56.08 ± 27.70 (N=19) 40.93 ± 20.93 (N=18) 15.16 (-1.30-31.62) 0.07

Secondary Analyses Method 1 = maximum dose   Treatment Group Mean±SD or no./N (%) Estimated Difference (95% CI) P Value Prednisone Alone Thymectomy +prednisone Time-weighted average prescribed AD prednisone dose (mg)a 59.3 ± 28.4 (N=56) 43.4 ± 23.1 (N=62) 16.0 (6.6-25.4) 0.001 Penalized time-weighted average AD prednisone dose (mg; max dose)a,b 72.9 ± 37.4 46.7 ± 24.9 (N=61) 26.3 (14.7-37.8) < 0.001 Penalized time-weighted AD average prednisone dose (mg; dose at start)a,c 68.1 ± 38.3 45.6 ± 24.3 22.5 (10.6-34.4) Time-weighted average MG Activities of Daily Livinga,d 3.41 ± 2.58 (N=55) 2.24 ± 2.09 1.17 (0.31-2.03) 0.008 MG ADL at month 12 3.33 ± 3.40 (N=54) 1.92 ± 2.73 (N=61) 1.42 (0.28-2.55) 0.01 MG ADL at month 24 3.11 ± 2.93 (N=53) 2.02 ± 2.78 (N=59) 1.10 (0.03-2.17) 0.04 MG ADL at month 36 2.69 ± 2.80 (N=51) 2.14 ± 2.92 (N=59) 0.55 (-0.53-1.63) 0.32 Azathioprine usee 28/58 (48) 11/65 (17) 31.4% (15.6%,47.1%) Plasma exchange usee 9/58 (16) 10/65 (15) 0.1% (-12.7%,12.9%) ~1 Intravenous immunoglobulin usee 23/58 (40) 22.7% (7%,38.3%) 0.005 Minimal Manifestation Statuse at month 12f 20/54 (37) 41/61 (67) 30.2% (12.7%-47.6%) at month 24f 20/53 (38) 39/59 (66) 28.4% (10.6%-46.2%) 0.003 at month 36f 24/51 (47) 39/58 (67) 20.2% (1.9%-38.5%) 0.03 Hospitalization for MG exacerbation Months 0-24: # of patientsg 17/60 (28) 6/66 (9) 19.2% (5.9%, 32.6%) 0.006 Cumulative daysh 26.4 ± 28.9 5.5 ± 2.9 0.004 Months 0-36: # of patientsg 22/60 (37) 27.6% (13.6%, 41.6%) <0.001 22.5± 27.1 8.7 ± 7.7 0.21 Method 1 = maximum dose Method 2 = Last dose prior to initiation of Prednison

Summary of Adverse Events † Prednisone Alone (N=60) Thymectomy+ Prednisone (N=66) Number of events 93 48 Patients having ≥1 event– no.(%) 33 (55) 25 (38)   Classification by patient no.(%) Life threatening 7 (12) 1 (2) Disability/Incapacity† 2 (3) 8 (12) Required medical or surgical intervention 5 (8) 9 (14) Death 0 (0) Complication due to thymectomy Not applicable Hospitalization 31 (52) 15 (23) Patients hospitalized for MG exacerbation 22 (37) 6 (9) Mean ± SD cumulative hospital days 19.2 ± 24.5 8.4 ± 8.6 Hospitalization by MEDRA codes Gastrointestinal disorders Hepatobiliary disorders Infections and infestations 4 (6) Injury, poisoning and procedure complications Metabolism and nutrition disorders Nervous system disorders Respiratory, thoracic and mediastinal disorders Surgical and medical procedures Vascular disorders Disability/incapacity etiologies: for prednisone alone group, worsening swallowing difficulties and myasthenia gravis; in thymectomy+prednisone group, osteoporotic thoracic fracture, ocular muscle involvement due to relapsing MG, post-thymectomy diaphragmatic hemiparesis, rib fracture, impending myasthenic crisis, Pott’s fracture, tear of left knee meniscus, and low back pain with possible stenosis. MEDRA denotes medical dictionary for regulatory activities, MG myasthenia gravis.

Treatment-associated symptoms Patients with ≥ 1 symptom p=0.22 over 0-12 months p=0.002 over 0-24 months p<0.001 over 0-36 months Mean no. of symptoms p=0.007 over 0-12 months p=0.02 over 0-24 months p<0.001 over 0-36 months Mean distress level (0-4, “not at all” to “extremely”) p=0.04 over 0-12 months p=0.003 over 0-24 months p=0.003 over 0-36 months Treatment-associated complications and SF-36 without significant differences

Thymic Histology from 46 Patients Randomized to ETTX   Grade 0 Grade 1 Grade 2 Grade 3 Grade 4 TFH (CD23+)a 15 18 6 5 2 Overall atrophy 1 7 14 Cortical atrophy 10 12 4 67% of specimens with TFH aThymic follicular hyperplasia was defined by counting CD23+ follicular dendritic cell networks and graded as: Grade 0, no follicles; Grade 1, follicles in ≤1/3 of thymic lobules; Grade 2, follicles in >1/3 and ≤ 2/3 of thymic lobules; Grade 3, follicles in >2/3 of thymic lobules; Grade 4, lymph node-like transformation of the thymus with >4 follicles per low power field (x50 magnification). One patient found to have WHO type B2 thymoma (excluded from above)

BioMG Analysis of thymus samples Standard operating and handling procedures working well Well suited for biomarker studies Grading system for lymphofollicular hyperplasia to be correlated with clinical data Marx et al. Ann NY Acad Sci 2012;1275:92

MGTX in Context Statements for non-thymomatous MG Cochrane Collaboration “There is no randomized controlled trial literature that allows meaningful conclusions about the efficacy of thymectomy on MG. Data from several class III observational studies suggest that thymectomy could be beneficial in MG. An RCT is needed…” Cea G, et al. Cochrane Library 2013;10:1-20 Intl MG Treatment Recommendations “In non-thymomatous MG, thymectomy is performed as an option to potentially avoid or minimize the dose or duration of immunotherapy, or if patients fail to respond to an initial trial of immunotherapy or have intolerable side effects from that therapy.” Sanders DB, Wolfe GI, Benatar M, et al. Neurology 2016; 87:419-425

MGTX Summary Class I evidence: ETTX has favorable impact on AChRAb+ generalized MG Clinical outcomes Corticosteroid exposure Adverse events Effect can be seen in <12 months Future investigations Predictive role of thymic histology BioMG: Genomic/proteomic profiles, biomarker assessments Wolfe GI et al. N Engl J Med 2016;375:511-522

Many Thanks! Study subjects: Trust in randomization; ≥3 yrs MGTX Investigators/DSMB March 2006 San Fran / Oxford