Treatment Approaches for Stage IIIa(N2) Lung Cancer Frank Detterbeck MD Thoracic Surgery, Yale University Yale Thoracic Oncology Program No conflicts to disclose with respect to this presentation

Disclosures Medela Research Grant (chest drainage device) Olympus Data Safety Monitoring board Covidien Research advisory board participation Lilly Lectures on 7th edition Lung Cancer Stage classification system in the past

“The answer to cancer” “If it bleeds… I can kill it!”

Fact: Survival is the same for ChemoRT and for Chemo/RT/Surgery

Bi- vs Tri-Modality Tmt of Stage III % Overall Survival Surgery (n = 167) RT (n = 165) Van Meerbeck JNCI 2007;99:442-50

Intergroup 0139/RTOG 9309 Overall Survival by Treatment Arms CT/RT/S 145/202 CT/RT 155/194 Logrank p = 0.24 Hazard ratio = 0.87 (0.70, 1.10) Odds Ratio at 5 yrs 0.63 (0.36-1.1) % Alive 25 50 75 100 Months from Randomization 12 24 36 48 60 Dead/Total

Stage III: Ch/RT vs Ch/RT/Surg Intergroup 0139 Is overall survival the same? Is one curve flatter than the other? 100 75 50 25 12 24 36 48 60 Is the glass: Half Full Half Empty? or

Fact: We have identified prognostic factors, (most are applicable only retrospectively) These ARE NOT predictive that surgery is beneficial

IIIa (N2) – It’s a Spectrum of Patients! Mediastinal Infiltration Discrete node enlargement Clinically occult N2 Schematic of types of patients included in studies using different treatment approaches Curative- Intent Chemo-RT Palliative Treatment (PS ≥ 2) Neoadjuvant + Surgery Primary Surgery (occult N2) Figure legend: Representative examples of infiltrative mediastinal nodal involvement (left) in which patients should generally be treated with chemotherapy (CT) and radiotherapy either concurrently fir good performance status patients (PS) or sequentially. Clinically occult N2 involvement, (far right; as suggested by the lack of uptake on PET scanning) that is discovered after thorough pre-operative staging and after surgical resection should be treated post-operatively with adjuvant cisplatin pbased chemotherapy, and radtiotherapy is reserved for when concern for local recurrence is high. Discrete N2 nodal involvement (center) MUST be confirmed pathologically before deciding on a treatment course that can include CT/RT or induction therapy followed by complete surgical resection. Incomplete resection should be avoided and patients in whom complete resecton is not feasible should be treated with CT/RT. IIIa (N2) – It’s a Spectrum of Patients! ↑ Performance Status ↑ Tumor Burden Stage III patient characteristics

Fate of Patients selected for Trimodality Tmt 402 Good-risk patients, N2 at EBUS/Mediastinoscopy Thought to be good candidates for Trimodality Tmt Died Lost R1,2 N2 at Restage Indctn Toxicity Progresed R0 Ref: Cerfolio Ann Thor Surg 2008;86:912-20

Fate of Patients selected for Trimodality Tmt 402 Good-risk patients, N2 at EBUS/Mediastinoscopy selected for Trimodality Tmt 5-Yr Survival (%) % of pts 100 10 11 3 16 8 Neg at Restage Ref: Cerfolio ATS 2008;86:912-20

Common Arguments for Surgery in N2 Selection Criteria Assessment of commonly cited arguments Summary: Justifica-tion for Surgery Pre-operatively identifiable? Prognostic value? Potential Flaw Defines treatment value? “Minimal” N2 Moderate Probably Out-of-context Unclear Single station Yes cN0,1 Non-bulky nodes Good surg risk - No Not Applic Downstaged Limited Landmark Shrinkage Subjective Lobectomy

Fact: Surgery causes Morbidity And Mortality

INT 0139: Pneumonectomy Pts vs Matched CT/RT Subset 100 75 / % Overall Survival 50 45% 36% / / / / / / 24% 22% ChRT / 25 / / ChRT/S logrank p = NS 12 24 36 48 60 Months

Op. Mortality of Neoadj  Pneumonectomy Study N (pneum) Ind Tmt % Mort Stamatis 02 127 Ch 7 Alifano 08 113 6 Doddoli 05 100 12 Martin 01 97 11 Van Schil 05 69 Matsubara 04 68 4 Mansour 07 60 Albain 05 54 Ch/RT 26 Gudbjartson 08 35 Daly 06 30 13 Perrot 05 27 Roberts 01 20 Average 8 Inclusion Criteria: Studies of ≥20 pts 2000-08

Are there any Answers?

Treatment of Stage III NSCLC Role of Trimodality is based on flawed arguments Retrospective selection Inappropriate application of subset results to the whole But: Trimodality and Bimodality are the same So if the patient really wants it, it is OK IF your periop morbid/mortality is minimized IF you can achieve a complete resection IF you can manage the distant disease