ESMO Magnitude of Clinical Benefit Scale for new anticancer drugs Roundtable on Immuno-Oncology Federal Parliament – Brussels, July 02nd 2015 ESMO Magnitude of Clinical Benefit Scale for new anticancer drugs Martine Piccart on behalf of ESMO Magnitude of Clinical Benefit Scale task force Institut Jules Bordet, Université Libre de Bruxelles Cette présentation est la propriété de l’ESMO www.esmo.org

New anticancer drugs hit the market at very high costs ! The average cost of recently approved anticancer drugs is ≈ 9.000 €/month The cost to gain one year of life In 1995 ≈ 50.000 € Today ≈ 200.000€ L. Salz – ASCO 2015

Value ≠ Benefit

Possible scenarios in outcome of pivotal, randomized phase III clinical trials Survival results No difference Small, statistically significant difference possible EMA registration of new treatment Large, statistically significant difference very likely EMA registration of new treatment Standard treatment New treatment

Different impacts of new drugs on patient outcome Standard treat A B C The new treatment controls the disease better initially, but the life of the pt is not extended. What about QoL ? New treat A B C The new treatment controls the disease better and the pt lives longer New treat A B C Is this extension measured in days, weeks or months ?

Why an ESMO Magnitude of Clinical Benefit Scale (ESMO-MCBS) ? Committed to promote high-quality, rational, responsible & affordable cancer care Recognizes the need for clear and unbiased statements regarding the magnitude of clinical benefit from new therapeutic approaches Wants to highlight treatments which bring substantial improvements to the duration of survival and/or the QoL of cancer patients Hopes The scale will facilitate access to important anticancer drugs all over Europe

Factors taken into account for ESMO-MCBS Overall survival, Progression free survival Quality of Life Magnitude of Clinically Benefit Toxicity Prognosis of the condition Costs Not analyzed in view of significant “Heterogeneity” across Europe

ESMO-MCBS substantial improvements Curative setting A & B or non-curative setting 5 & 4 These drugs should ideally be accessible to all European citizens Curative Non-curative 5 4 3 2 1 A B C

Evaluation form 1: for adjuvant and other treatments with curative intent Mark with X if relevant Grade A >5% improved survival at ≥ 3 years follow-up Improvement in DFS alone (primary endpoint) (HR < 0.65) in studies without mature survival data Grade B ≥ 3% but ≤ 5% improvement at ≥ 3 years follow-up Improvement in DFS alone (primary endpoint) (HR 0.65 - 0.8) without mature survival data Non inferior OS or DFS with reduced treatment toxicity or improved QoL (with validated scales) Non inferior OS or DFS with reduced treatment cost as reported study outcome (with equivalent outcomes and risks) Grade C < 3% improvement at ≥ 3 years follow-up Improvements in DFS alone (primary endpoint) (HR > 0.8) in studies without mature survival data

Evaluation form 1: for adjuvant and other treatments with curative intent Mark with X if relevant Grade A >5% improved survival at ≥ 3 years follow-up Improvement in DFS alone (primary endpoint) (HR < 0.65) in studies without mature survival data Grade B ≥ 3% but ≤ 5% improvement at ≥ 3 years follow-up Improvement in DFS alone (primary endpoint) (HR 0.65 - 0.8) without mature survival data Non inferior OS or DFS with reduced treatment toxicity or improved QoL (with validated scales) Non inferior OS or DFS with reduced treatment cost as reported study outcome (with equivalent outcomes and risks) Grade C < 3% improvement at ≥ 3 years follow-up Improvements in DFS alone (primary endpoint) (HR > 0.8) in studies without mature survival data

ESMO-MCBS distinctions: for treatment with non-curative intent Primary endpoint OS PFS or TTP Other than OS or PFS Median with standard therapy ≤ 1 year > 1 year Median with standard therapy ≤ 6 months > 6 months

A. Application of the scale to new drugs for poor prognosis diseases Non curative setting

Evaluation form 2a: treatments with non-curative intent, primary endpoint OS IF median OS with the standard treatment is ≤ 1 year Mark with X if relevant Grade 4 HR ≤ 0.65 AND Gain ≥ 3 months Increase in 2 year survival alone ≥ 10% Grade 3 HR ≤ 0.65 AND Gain 2.5-2.9 months Increase in 2 year survival alone 5- <10% Grade 2 HR > 0.65-0.70 OR Gain 1.5-2.4 months Increase in 2 year survival alone 3- <5% Grade 1 HR > 0.70 OR Gain < 1.5 month Increase in 2 year survival alone < 3%

Application of the scale for diseases with a medium prognosis Non curative setting

Evaluation form 2a: treatments with non-curative intent, primary endpoint OS IF median OS with the standard treatment is > 1 year Mark with X if relevant Grade 4 HR ≤ 0.70 AND Gain ≥ 5 months Increase in 3 year survival alone ≥ 10% Grade 3 HR ≤ 0.70 AND Gain 3-4.9 months Increase in 3 year survival alone 5- <10% Grade 2 HR > 0.70-0.75 OR Gain 1.5-2.9 months Increase in 3 year survival alone 3- <5% Grade 1 HR > 0.75 OR Gain < 1.5 month Increase in 3 year survival alone <3%

(highest grade scored) Evaluation form 2a: treatments with non-curative intent, primary endpoint OS Preliminary magnitude of clinical benefit grade (highest grade scored) Step 1 4 3 2 1 Step 2 Assessment QoL & grade 3-4 toxicities Does secondary endpoint QoL show improvement Are there statistically significantly fewer grade 3-4 toxicities impacting daily well-being* *not including alopecia, myelosuppression, but rather chronic nausea, diarrhea, fatigue, etc. Adjustment: Upgrade 1 level if improved QoL or less toxicity or is shown Step 3 Final adjusted magnitude of clinical benefit grade 5 4 3 2 1

ESMO-MCBS distinctions: for treatment with non-curative intent Primary endpoint OS PFS or TTP Other than OS or PFS Median with standard therapy ≤ 1 year > 1 year Median with standard therapy ≤ 6 months > 6 months

Evaluation form 2b: treatments with non-curative intent, primary endpoint PFS or TTP Studies with median PFS with standard treatment > 6 months Mark with X if relevant Grade 3 HR ≤ 0.65 AND Gain ≥ 3 months Grade 2 HR ≤ 0.65 BUT Gain < 3 months Grade 1 HR > 0.65

(highest grade scored) Evaluation form 2b: treatments with non-curative intent, primary endpoint PFS or TTP Preliminary magnitude of clinical benefit grade (highest grade scored) Step 1 3 2 1 Step 2 Look at OS : if improved, go to the OS scale; if not improved, go to steps 2 and 3 Toxicity and QoL adjustment when only a PFS improvement

Evaluation form 2b: treatments with non-curative intent, primary endpoint PFS or TTP Toxicity assessment (adverse effect criterion) Is the new treatment associated with a statistically significant incremental rate of: «toxic» death > 2% cardiovascular ischemia > 2% hospitalization for «toxicity» > 10% excess rate of severe CHF > 4% grade 3 neurotoxicity > 10% severe other irreversible or longlasting toxicity > 2% please specify: (Incremental rate refers to the comparison versus standard therapy in the control arm) Step 2 Mark with X if relevant Assessment QoL & grade 3-4 toxicities Step 3 Was QoL measured as a secondary outcome Does secondary endpoint QoL show improvement Are there statistically significantly fewer grade 3-4 toxicities impacting daily well-being* *not including alopecia, myelosuppression, but rather chronic nausea, diarrhea, fatigue, etc.

Evaluation form 2b: treatments with non-curative intent, primary endpoint PFS or TTP Step 4 Final Adjustments Downgrade 1 level if ≥ 1 of above incremental toxicities Upgrade 1 level if > QoL or if <grade 3-4 toxicities that bother patients When OS as 2nd endpoint is improved, it prevails, score according to form 2a Downgrade 1 level if the drug ONLY leads to improved PFS and QoL assessment does not demonstrate improved QoL Final, toxicity and QoL adjusted, magnitude of clinical benefit grade 4 3 2 1 Highest grade that can be achieved is grade 4

Field testing Breast Cancer Medication Trial Setting Primary outcome PFS control PFS gain PFS HR OS control OS gain OS HR QoL ESM0-MCBS Chemo +/- trastuzumab HERA (Neo)Adjuvant HER-2 positive tumors DFS 2 y DFS 77.4% 8.4% 0.54 (0.43-0.67) A T-DM1 vs capecitabine + lapatinib EMILIA 2nd line metastatic after trastuzumab failure PFS & OS 6.4 m 3.2 m 0.65 (0.55-0.77) 25 m 6.8 m 0.68 (0.55-0.85) Later deterioration 5 Trastuzumab + chemo +/- pertuzumab CLEOPATRA 1st line metastatic PFS 12.4 m 6 m 0.62 (0.52-0.84) 40.8 m 15.7 m (0.56-0.84) ~ 4 Lapatinib +/- trastuzumab EGF 104900 3rd line metastatic 2 m 1 m 0.73 (0.57-0.93) 9.5 m 4.5 m 0.74 (0.57-0.97) Capecitabine +/- lapatinib Geyer, 2006 4.4 m 4 m 0.49 (0.34-0.71) NS 3 Eribulin vs other chemo EMBRACE 3rd line metastatic after anthracycline & taxane OS 10.6 m 2.5 m 0.81 (0.66-0.99) 2 Paclitaxel +/- bevacizumab Miller, 2007 5.9 m 5.8 m 0.6 (0.51-0.70) Exemestane +/- everolimus BOLERO-2 Metastatic after failure aromatase inhibitor+PFS >6 m 4.1 m 6.5 m 0.43 (0.36-0.54)

Field testing Lung Cancer (1) Medication Trial Setting Primary outcome PFS control PFS gain PFS HR OS control OS HR QoL Toxicity ESM0-MCBS Erlotinib vs carboplatin gemcitabine OPTIMEL, CTONG-0802 1st line stage 3b/4 non-squamous + EGFR mutation 4.6 m 8.5 m 0.16 (0.10-0.26) 12% < serious adverse events 4 Erlotinib vs Pt-based chemo doublet EURTAC PFS, crossover allowed 5.2 m 4.5 m 0.37 (0.25-0.54) 19.5 m NS 15% < severe adverse reactions Gefitinib vs carboplatin + paclitaxel IPASS 6.3 m 3.3 m 0.48 (0.34-0.67) < toxicity Afatinib vs cisplatin + pemetrexed LUX Lung 3 6.9 m 4.2 m 0.58 (0.43-0.78) Del19/L858R 6.7 m 0.47 (0.34-0.65) Crizotinib vs chemo Shaw 2013 1st line stage 3b/4 non-squamous + ALK mutation 3.0 m 4.7 m 0.49 (0.37-0.64) 1% > toxic death Crizotinib vs cisplatin + pemetrexed Solomon 2014 7.0 m 3.9 m 0.45 (0.35-0.60)

Conclusions and next steps ESMO-MCBS offers an objective and reproducible approach to grade drugs and To determine which drugs are immediately required for European citizens Relevance clear based on ESMO Anti-neoplastic medicines survey To include findings in guidelines To support clinical decision making counseling patients editorial decisions and commentaries ESMO-MCBS v1 will be published in Annals of Oncology, early release http://annonc.oxfordjournals.org/lookup/doi/10.1093/annonc/mdv249 Version 1 A «lively» instrument needing regular updates by ESMO committee As more mature data of clinical trials become available, allowing fine tuning for efficacy and toxicity New drugs and new indications will become available Adaptations of the grading of the scale may be required

Acknowledgments Task Force members Martine Piccart, Co-chair Richard Sullivan Nathan Cherny Urania Dafni Martijn Kerst Alberto Sobrero Christoph Zielinski ESMO ex Board ESMO Staff: Keith McGregor and Nicola Latino Numerous people who helped testing the scale

BACK-UPS

Differences in access to relevant new anticancer drugs in Europe Differences between countries in: drug related health care expenditures drug prices access time to drugs after approval by EMA Sometimes lack of drug supply in “countries with cheaper drugs” due to parallel import to “countries where the drug is more expensive”. Unequal access within some countries: sometimes (co)-payment of the drug costs by patients required

ESMO-MCBS for solid tumors was developed with “Snowball” method

“Snowball” method + Simulation scenarios Previous work of Task Force Members 1st draft scale Biostatisticians 1st ESMO faculty field testing 2nd draft scale Applied in wide range of settings by Task Force & invited experts 2nd ESMO faculty field testing Final Scale 13 drafts scale Feedback Integration work Sobrero on role HR, prognosis & absolute differences in data interpretation + Sobrero A et al. J Clin Oncol 2009 Sobrero A et al Clin Cancer Res 2015

Underlying Premises ESMO-MCBS Cure takes precedence over deferral of death Direct endpoints such as survival and QoL take precedence over surrogates such as PFS or RR DFS in curative disease is a more valid surrogate than PFS or RR in non-curative disease Interpretation of the evidence for benefit derived from surrogate outcomes (such as PFS) may be influenced by secondary outcome data

3 Rules, #1 ESMO-MCBS Data derived from comparative research: Priority: Strong level of evidence from large phase III studies, > lesser level of evidence from comparative cohort studies or randomized phase II studies Careful analyses “control arm” and identification of endpoints. Subgroup analysis.    preplanned in ESMO-MCBS when ≤ 3 subgroups defined «a priori»: benefit in a subgroup for the primary endpoint can be «scaled», provided adjusted for multiple comparisons unplanned not in ESMO-MCBS considered hyposthesis generating  

Example: for threshold set at HR ≤ 0.70 3 Rules, #2 ESMO-MCBS More than one outcome may be applicable For a required HR, not the point estimate but the lower limit of the 95% CI is used to take into account the variability of the estimate Example: for threshold set at HR ≤ 0.70 it is the lower limit of the 95%CI which has to be ≤ 0.70 0.71 0.78 0.86 Trial X 0.65 0.76 0.89 Trial X does not qualify Trials Y and Z do qualify Trial Y 0.58 0.69 0.82 Trial Z HR 0.5 0.7 1.0

3 Rules, #3 ESMO-MCBS Check for: Ni Check for: indicators of severe toxicity or reduced grade 3-4 toxicity that bothers patients global QoL advantage using validated scale Report final adjusted grade taken into account toxicity and QoL when applicable toxicity quality of life

Forms ESMO-MCBS Curative Setting → Evaluation form 1 Non-curative setting → Evaluation form 2 a, b, c 2a: primary endpount OS 2b: primary endpoint PFS or TTP 2c: other primary endpoint On top of each form

Field testing Melanoma (2) version light Medication Trial Setting ESM0-MCBS Dabrafenib + trametinib vs vemurafenib Robert 2015 1st line unresectable or metastatic + BRAF V600E mutation 4* Vemurafenib +/- cobimetinib Larkin 2014 Dacarbazine +/- nevolumab 4 Dacarbazine +/- ipilimimab Robert 2011 Maio 2015 1st line metastatic 3 * immature survival data