Studies grouped by E-lesion prognostic influence
| No . | Year . | Author . | Reference . | Impact of E-lesions on prognosis and noted covariates . | Treatment . | OS . | DFS/PFS/FFTF . | Relapse rate . | CR/duration/other outcomes . |
|---|---|---|---|---|---|---|---|---|---|
| A. Studies concluding E-lesions are a poor prognostic factor under nuanced circumstances | |||||||||
| 2 | 1977 | Levi et al | 22 | E-lesions were associated with significantly more relapses, shorter remission duration, and worse OS in EFRT-only group but not CMT group. Strong association between lung E-lesions, moderate to large mediastinal masses, and subsequent marginal recurrences. | RT | 4-year OS worse with E-lesion RT-only group: nodal ∼87% vs E-lesion ∼56% (P ≤ .01) | Significantly shorter remission duration for the patients with “E”-stage disease when extended field irradiation was the initial therapy (P < .002) | Relapse after 5 years: 29% nodal disease (14/48) vs 82% with E-lesions (9/11) | 3-year CR duration: nodal ∼71% vs E-lesion ∼18% |
| CMT | 5-year OS not significantly different with CMT: nodal ∼100% vs E-lesion ∼97% (P > .10) | No significant difference between the 2 patient groups when initial therapy was CMT (P > .35) | Relapse after 5 years: nodal disease (6%, 2/36) vs E-lesions (14%, 1/7) | 5-year CR duration: nodal ∼95% vs E-lesion ∼86% | |||||
| 1982 | Wiernik and Slawson | 23 | RT | 12-year OS: nodal ∼90% vs E-lesion ∼60% (P < .03) | 12-year DFS of original 115 pts: Nodal 78% vs E-lesion 28% (P = .002) | nd | 12-year CR duration: Nodal ∼83% vs E-lesion ∼38% (p = 0.001) | ||
| CMT | 12-year OS: nodal ∼98% vs E-lesion ∼93% (P > .4) | 12-year DFS of original 115 pts: nodal 84% vs E-lesion 94% (P = .002) | nd | 12-year CR duration: nodal ∼97% vs E-lesion ∼90% (P > .3) | |||||
| 4 | 1984 | Zagars and Rubin | 56 | E-lesions associated with significantly more relapses in stage IIA treated with RT alone. In 2 cases, mediastinal masses may have driven poor prognosis. In small portion with CMT, E-lesions did not show negative prognostic implication. | RT vs CMT | nd | nd | Significantly more relapses, P < .05 | nd |
| 15 | 2003 | Hodgson et al | 39 | 5-year OS: no significant effect | CMT | nd | 5-year DFS: no E-lesion (84%) vs chest wall (59%) (P = .016) | nd | 5-year cause-specific survival: No E (94%) vs chest wall (86%) (P = .009) |
| Chest wall significantly worse; 5-year CSS and DFS: lung extension no significant effect | CMT | 5-year OS: 90% no E-lesion, 82% chest-wall invasion (P = .095), 88% lung invasion (P = .386) | 5-year DFS: no E (84%) vs lung invasion (80%) (P = .47) | nd | 5-year cause-specific survival: No E (94%) vs lung invasion 92% (P = .25) | ||||
| 17 | 2004 | Vassilakopoulos et al | 41 | In A(E)BVD subgroup (but not all patients), E disease was an independent predictor of poorer 10-year FFS. E disease status did not impact 10-year OS. | CMT | nd | 10-year FFS: A(E)BVD subgroup, no (87%) vs (73%) (P = .03) | nd | FFTF in patients who achieved CR, CR, or VGPR and received low-dose RT, E-lesion prognostic factor P < .001 in all treatment groups. |
| CMT | 10-year OS: all patients, no E (86%) vs E (94%) (P = .30); A(E)BVD only, no E (93%) vs E (100%) (P = .37) | 10-year FFS: all patients, no E (85%) vs E (75%) (P = .11) | nd | nd | |||||
| B. Studies concluding E-lesions are not a prognostic factor | |||||||||
| 3 | 1982 | Hoppe et al | 25 | E-lesions were not a significant factor for OS or FFR within treatment groups. | RT | 5-year OS: E-lesions, 100% vs all patients, 96% | 5-year FFR: E-lesions, 78% vs all patients, 79% | nd | nd |
| CMT | 5-years OS: E-lesions, 88% vs all patients, 92% | 5-year FFR: E-lesions, 90% vs all patients, 87% | nd | nd | |||||
| 1986 | Crnkovich et al | 26 | E-lesions were not a significant factor for 10-year OS and FFR between treatment groups. | RT | 10-year OS: RT arm, 87% vs all patients, 87% | 10-year FFR: RT arm, 70% vs all patients, 71% | nd | nd | |
| CMT | 10-year OS: E-lesions, 72% vs all patients, 74% | 10-year FFR: CMT arm, 71% vs all patients, 79% | nd | nd | |||||
| 6 | 1985 | Leslie et al | 35 | When analyzed by mediastinal size, E-lesions did not influence FFR or OS. Among patients with E-stage disease, 11 of 25 had B-symptoms and 13 of 25 had bulk disease. | RT or CMT | 10-year OS for large mediastinal adenopathy with E vs without E: 85% vs 81% | 10-year FFR for large mediastinal adenopathy with E vs without E: 62% vs 58% | nd | nd |
| 10-year OS for small mediastinum with E vs without E: 80 vs 84% | 10-year FFR for small mediastinum with E vs without E: 85% vs 81% | nd | nd | ||||||
| 7 | 1989 | Loeffler et al | 28∗ | E-lesions had no prognostic significance for CR rate, FFTF, OS, or FFP. Six of the patients with stage IIAE disease and 5 of the patients with stage IIBE disease also had MLT. | CMT | 3-year OS: no influence of E-lesions | FFTF: all patients vs E, 20% vs 25%, not significant | nd | CR rate: all patients vs E, 83% vs 81% |
| 1997 | Loeffler et al | 27∗ | Despite E-lesions not routinely irradiated (lung and pleura), there were 100% local CR and no relapses after chemotherapy (median follow-up, 6.5 y). | CMT | nd | nd | No relapses after chemotherapy; median follow-up, 6.5 y | 100% local CR; median follow-up, 6.5 y | |
| 8 | 1989 | Leopold et al | 36∗ | E-lesions had no significant effect on 12-year relapse rate or survival within treatment groups. | RT | 12-year OS: no effect in different treatment groups | nd | Relapse rate not significantly different in treatment groups | nd |
| CMT | 12-year OS: no effect in different treatment groups | nd | Relapse rate not significantly different in treatment groups | nd | |||||
| 10 | 1999 | Shah et al | 37 | E-lesions had no significant influence on relapse rate, OS, or EFS. | RT | 10-year OS: no influence | 10-year EFS: no influence | nd | nd |
| 14 | 2003 | Glimelius et al | 38 | Not significant in 10-year DFS and HL-specific survival | CMT | nd | 10-year DFS and HL-specific survival not significant | nd | nd |
| 22 | 2013 | Song et al | 50 | E-lesion not a significant factor for PFS and OS at 45 months. | All chemo, some CMT | Hazard ratio for OS if E-lesion: 2.581; 95% CI, 0.916-7.273; P = .073 | Hazard ratio for PFS if E-lesion: 1.762; 95% CI, 0.661-4.698; P = .258 | nd | nd |
| C. Studies concluding E-lesions are a poor prognostic factor | |||||||||
| 5 | 1985 | Zittoun et al | 34 | Patients with stage IIE disease had significantly worse 5-year DFS than defined low-risk groups. | CMT | nd | 5-year DFS: significantly worse for patients with stage IIE (significance not provided). | nd | nd |
| 11 | 2000 | Franklin et al | 18∗ | Patients with E-lesions, especially IIBE and IIIE, had a poor prognosis at 5 years. E-lesions were a significant poor prognostic factor beyond IPS. | CMT | nd | DFS: Cox regression with IPS and additional factors after 5 years: stage IIBE and IIIE significant in addition to IPS (P = .017), hazard ratio: 2.62; all E-lesions HR 1.54; P = .086 | nd | nd |
| 2002 | Sieber et al | 29∗ | 7-year FFTF and CR rate significantly worse in both treatment arms. | CMT | nd | 7-year FFTF: significantly worse (P = .015) | nd | CR duration significantly worse COPP/ABVD arm, 86% vs 94% (P = .011) and COPP/ABV/IMEP arm 82% vs 86% (P ≤ .001) | |
| 12 | 2001 | Ruhl et al | 33 | E-lesions were a significant risk factor for both progressive disease and relapse. | All chemo, some CMT | nd | Risk factor for progressive disease (P ≤ .002); median follow-up time, 38 mo | Risk factor for relapse (P < .002); median follow-up time, 38 mo | nd |
| 18 | 2005 | Gisselbrecht et al | 42 | E-lesions associated with significantly worse OS at 42 mo. Authors hypothesize E disease may be surrogate for bulky mediastinal disease. | All chemo, most CMT | Multivariate analysis stage–adjusted prognostic index 42-mo OS RR: 1.2 (P = .008) | nd | nd | nd |
| 20 | 2011 | Wirth et al | 48 | Significant factor for worse OS and FFS at 5 y | CMT | 5-year OS: no E vs E, 97% (95% CI, 93-100) vs 67% (95% CI, 20-90); P = .0005. | 5-year PFS: no E vs E: 92% (95% CI, 86-96) vs 50% (95% CI, 11-80); P = .0002) | nd | Remained significant factor in multivariate analysis. |
| 21 | 2012 | Gobbi et al | 49∗ | For early, unfavorable disease, presence of E-lesions was the only statistically important predictor of early treatment failure beyond relative tumor burden. | CMT | nd | nd | nd | Early treatment failure predictor in addition to relative tumor burden: E-lesions coef: 0.846; risk, 2.329; P = .0329 |
| 23 | 2014 | Laskar et al | 44∗ | Significant factor for worse 10-year PFS and OS | CMT | 10-year OS: no E (96%) vs E (0%) (P = .01) | 10-year PFS: E significantly worse (P = .037) | nd | nd |
| No . | Year . | Author . | Reference . | Impact of E-lesions on prognosis and noted covariates . | Treatment . | OS . | DFS/PFS/FFTF . | Relapse rate . | CR/duration/other outcomes . |
|---|---|---|---|---|---|---|---|---|---|
| A. Studies concluding E-lesions are a poor prognostic factor under nuanced circumstances | |||||||||
| 2 | 1977 | Levi et al | 22 | E-lesions were associated with significantly more relapses, shorter remission duration, and worse OS in EFRT-only group but not CMT group. Strong association between lung E-lesions, moderate to large mediastinal masses, and subsequent marginal recurrences. | RT | 4-year OS worse with E-lesion RT-only group: nodal ∼87% vs E-lesion ∼56% (P ≤ .01) | Significantly shorter remission duration for the patients with “E”-stage disease when extended field irradiation was the initial therapy (P < .002) | Relapse after 5 years: 29% nodal disease (14/48) vs 82% with E-lesions (9/11) | 3-year CR duration: nodal ∼71% vs E-lesion ∼18% |
| CMT | 5-year OS not significantly different with CMT: nodal ∼100% vs E-lesion ∼97% (P > .10) | No significant difference between the 2 patient groups when initial therapy was CMT (P > .35) | Relapse after 5 years: nodal disease (6%, 2/36) vs E-lesions (14%, 1/7) | 5-year CR duration: nodal ∼95% vs E-lesion ∼86% | |||||
| 1982 | Wiernik and Slawson | 23 | RT | 12-year OS: nodal ∼90% vs E-lesion ∼60% (P < .03) | 12-year DFS of original 115 pts: Nodal 78% vs E-lesion 28% (P = .002) | nd | 12-year CR duration: Nodal ∼83% vs E-lesion ∼38% (p = 0.001) | ||
| CMT | 12-year OS: nodal ∼98% vs E-lesion ∼93% (P > .4) | 12-year DFS of original 115 pts: nodal 84% vs E-lesion 94% (P = .002) | nd | 12-year CR duration: nodal ∼97% vs E-lesion ∼90% (P > .3) | |||||
| 4 | 1984 | Zagars and Rubin | 56 | E-lesions associated with significantly more relapses in stage IIA treated with RT alone. In 2 cases, mediastinal masses may have driven poor prognosis. In small portion with CMT, E-lesions did not show negative prognostic implication. | RT vs CMT | nd | nd | Significantly more relapses, P < .05 | nd |
| 15 | 2003 | Hodgson et al | 39 | 5-year OS: no significant effect | CMT | nd | 5-year DFS: no E-lesion (84%) vs chest wall (59%) (P = .016) | nd | 5-year cause-specific survival: No E (94%) vs chest wall (86%) (P = .009) |
| Chest wall significantly worse; 5-year CSS and DFS: lung extension no significant effect | CMT | 5-year OS: 90% no E-lesion, 82% chest-wall invasion (P = .095), 88% lung invasion (P = .386) | 5-year DFS: no E (84%) vs lung invasion (80%) (P = .47) | nd | 5-year cause-specific survival: No E (94%) vs lung invasion 92% (P = .25) | ||||
| 17 | 2004 | Vassilakopoulos et al | 41 | In A(E)BVD subgroup (but not all patients), E disease was an independent predictor of poorer 10-year FFS. E disease status did not impact 10-year OS. | CMT | nd | 10-year FFS: A(E)BVD subgroup, no (87%) vs (73%) (P = .03) | nd | FFTF in patients who achieved CR, CR, or VGPR and received low-dose RT, E-lesion prognostic factor P < .001 in all treatment groups. |
| CMT | 10-year OS: all patients, no E (86%) vs E (94%) (P = .30); A(E)BVD only, no E (93%) vs E (100%) (P = .37) | 10-year FFS: all patients, no E (85%) vs E (75%) (P = .11) | nd | nd | |||||
| B. Studies concluding E-lesions are not a prognostic factor | |||||||||
| 3 | 1982 | Hoppe et al | 25 | E-lesions were not a significant factor for OS or FFR within treatment groups. | RT | 5-year OS: E-lesions, 100% vs all patients, 96% | 5-year FFR: E-lesions, 78% vs all patients, 79% | nd | nd |
| CMT | 5-years OS: E-lesions, 88% vs all patients, 92% | 5-year FFR: E-lesions, 90% vs all patients, 87% | nd | nd | |||||
| 1986 | Crnkovich et al | 26 | E-lesions were not a significant factor for 10-year OS and FFR between treatment groups. | RT | 10-year OS: RT arm, 87% vs all patients, 87% | 10-year FFR: RT arm, 70% vs all patients, 71% | nd | nd | |
| CMT | 10-year OS: E-lesions, 72% vs all patients, 74% | 10-year FFR: CMT arm, 71% vs all patients, 79% | nd | nd | |||||
| 6 | 1985 | Leslie et al | 35 | When analyzed by mediastinal size, E-lesions did not influence FFR or OS. Among patients with E-stage disease, 11 of 25 had B-symptoms and 13 of 25 had bulk disease. | RT or CMT | 10-year OS for large mediastinal adenopathy with E vs without E: 85% vs 81% | 10-year FFR for large mediastinal adenopathy with E vs without E: 62% vs 58% | nd | nd |
| 10-year OS for small mediastinum with E vs without E: 80 vs 84% | 10-year FFR for small mediastinum with E vs without E: 85% vs 81% | nd | nd | ||||||
| 7 | 1989 | Loeffler et al | 28∗ | E-lesions had no prognostic significance for CR rate, FFTF, OS, or FFP. Six of the patients with stage IIAE disease and 5 of the patients with stage IIBE disease also had MLT. | CMT | 3-year OS: no influence of E-lesions | FFTF: all patients vs E, 20% vs 25%, not significant | nd | CR rate: all patients vs E, 83% vs 81% |
| 1997 | Loeffler et al | 27∗ | Despite E-lesions not routinely irradiated (lung and pleura), there were 100% local CR and no relapses after chemotherapy (median follow-up, 6.5 y). | CMT | nd | nd | No relapses after chemotherapy; median follow-up, 6.5 y | 100% local CR; median follow-up, 6.5 y | |
| 8 | 1989 | Leopold et al | 36∗ | E-lesions had no significant effect on 12-year relapse rate or survival within treatment groups. | RT | 12-year OS: no effect in different treatment groups | nd | Relapse rate not significantly different in treatment groups | nd |
| CMT | 12-year OS: no effect in different treatment groups | nd | Relapse rate not significantly different in treatment groups | nd | |||||
| 10 | 1999 | Shah et al | 37 | E-lesions had no significant influence on relapse rate, OS, or EFS. | RT | 10-year OS: no influence | 10-year EFS: no influence | nd | nd |
| 14 | 2003 | Glimelius et al | 38 | Not significant in 10-year DFS and HL-specific survival | CMT | nd | 10-year DFS and HL-specific survival not significant | nd | nd |
| 22 | 2013 | Song et al | 50 | E-lesion not a significant factor for PFS and OS at 45 months. | All chemo, some CMT | Hazard ratio for OS if E-lesion: 2.581; 95% CI, 0.916-7.273; P = .073 | Hazard ratio for PFS if E-lesion: 1.762; 95% CI, 0.661-4.698; P = .258 | nd | nd |
| C. Studies concluding E-lesions are a poor prognostic factor | |||||||||
| 5 | 1985 | Zittoun et al | 34 | Patients with stage IIE disease had significantly worse 5-year DFS than defined low-risk groups. | CMT | nd | 5-year DFS: significantly worse for patients with stage IIE (significance not provided). | nd | nd |
| 11 | 2000 | Franklin et al | 18∗ | Patients with E-lesions, especially IIBE and IIIE, had a poor prognosis at 5 years. E-lesions were a significant poor prognostic factor beyond IPS. | CMT | nd | DFS: Cox regression with IPS and additional factors after 5 years: stage IIBE and IIIE significant in addition to IPS (P = .017), hazard ratio: 2.62; all E-lesions HR 1.54; P = .086 | nd | nd |
| 2002 | Sieber et al | 29∗ | 7-year FFTF and CR rate significantly worse in both treatment arms. | CMT | nd | 7-year FFTF: significantly worse (P = .015) | nd | CR duration significantly worse COPP/ABVD arm, 86% vs 94% (P = .011) and COPP/ABV/IMEP arm 82% vs 86% (P ≤ .001) | |
| 12 | 2001 | Ruhl et al | 33 | E-lesions were a significant risk factor for both progressive disease and relapse. | All chemo, some CMT | nd | Risk factor for progressive disease (P ≤ .002); median follow-up time, 38 mo | Risk factor for relapse (P < .002); median follow-up time, 38 mo | nd |
| 18 | 2005 | Gisselbrecht et al | 42 | E-lesions associated with significantly worse OS at 42 mo. Authors hypothesize E disease may be surrogate for bulky mediastinal disease. | All chemo, most CMT | Multivariate analysis stage–adjusted prognostic index 42-mo OS RR: 1.2 (P = .008) | nd | nd | nd |
| 20 | 2011 | Wirth et al | 48 | Significant factor for worse OS and FFS at 5 y | CMT | 5-year OS: no E vs E, 97% (95% CI, 93-100) vs 67% (95% CI, 20-90); P = .0005. | 5-year PFS: no E vs E: 92% (95% CI, 86-96) vs 50% (95% CI, 11-80); P = .0002) | nd | Remained significant factor in multivariate analysis. |
| 21 | 2012 | Gobbi et al | 49∗ | For early, unfavorable disease, presence of E-lesions was the only statistically important predictor of early treatment failure beyond relative tumor burden. | CMT | nd | nd | nd | Early treatment failure predictor in addition to relative tumor burden: E-lesions coef: 0.846; risk, 2.329; P = .0329 |
| 23 | 2014 | Laskar et al | 44∗ | Significant factor for worse 10-year PFS and OS | CMT | 10-year OS: no E (96%) vs E (0%) (P = .01) | 10-year PFS: E significantly worse (P = .037) | nd | nd |
A(E)BVD, adriamycin(epirubicin)/bleomycin/vinblastine/dacarbazine; chemo, chemotherapy; CI, confidence interval; CMT, combined modality therapy; coef, coefficient; COPP, cyclophosphamide, oncovin, prednison, procarbazin; CR, complete remission; CSS, cause specific survival; DFS, disease-free survival; E, epirubicin; EFRT, extended field radiotherapy; EFS, event-free survival; FFP, failure free progression; FFS, freedom from progression; FFTF, freedom from treatment failure; FFR, freedom from first relapse; HR, high risk; IPS, international prognostic score; MLT, large mediastinal tumor; nd, not defined; pts, patients; PFS, progression-free survival; RR, relative risk; RT, radiotherapy; VGPR, very good partial remission.
Study used elevated risk–based criteria for inclusion