Patients and tumor characteristics (Table 1)
Table 1
Variable | LD-SCCE, Number (%), n = 29 | ED-SCCE, Number (%), n = 27 |
Patient characteristics |
1.Age(in years) Median Interquartile range(IQR) | 58 50.5–66.5 | 57 51–64 |
2.Sex Male Female | 18(62.1%) 11(37.9%) | 15 (55.6%) 12(44.4%) |
3.Comorbidity Yes No | 14(48.3%) 15(51.7%) | 10(37.1%) 17(62.9%) |
4.ECOG PS 0–1 ≥ 2 | 28(96.5%) 1(3.5%) | 21(77.8%) 6(22.2%) |
5. Addiction* Tobacco chewer Smoker Alcohol None | 14(48.3%) 11(37.9%) 2(6.9%) 4(13.8%) | 11(40.7%) 12(44.4%) 5(18.5%) 4(14.8%) |
Tumor characteristics |
6.Location Cervical + upper Middle Lower + GEJ | 4(13.8%) 18(62.1%) 7(24.1%) | 3(11.1%) 13(48.1%) 11(40.7%) |
7.Histology Pure small cell carcinoma Mixed histology | 26(89.7%) 3(10.3%) | 25(92.6%) 2(7.4%) |
9. Median length in cm | 6.2 | 7.4 |
10.Skip lesion Yes No | 1(3.4%) 28(96.6%) | 2(7.4%) 25(92.6%) |
11.Regional lymph node involvement Yes No | 26(89.7%) 3(10.3%) | 26(96.3%) 1(3.7%) |
12.Metastatic sites Liver Non-regional lymph node Lungs Bone Bone marrow Brain | NA NA NA NA NA NA | 17(62.9%) 13(48.1%) 7(25.9%) 6(22.2%) 5(18.5%) 1(3.7%) |
13.Number of metastatic sites 1 2 ≥ 3 | NA NA NA | 13(48.1%) 7(25.9%) 7(25.9%) |
Laboratory parameters Median (IQR) Median (IQR) |
14.Hemoglobin in gm/dl | 12.3 (11.3–14.3) | 12.6 (11.6–14.1) |
15.Albumin in gm/dl | 4.1 (3.8–4.3) | 3.9 (3.6–4.2) |
16.Serum alkaline phosphatasein U/L | 98 (81–113) | 112 (87-148.5) |
17.LDH in U/L | 170 (152-196.5) | 328 (203–689) |
Abbreviations
LD-SCCE
Limited stage disease-small cell carcinoma esophagus
ED-SCCE
Extensive stage disease-small cell carcinoma esophagus
ECOG PS
Eastern Co-operative Oncology Group performance status
GEJ
Gastro-esophageal junction
Of 8342 patients with esophageal malignancies, 59(0.7%) had SCCE. No clinical details could be retrieved for 3 SCCE patients and they were therefore excluded. We included 56 patients in the study. Dysphagia (54 cases, 96.4%) was the commonest presenting symptom with a median duration of 2 months before presentation. Two patients (3.6%) had a family history of malignancy.
The tumors commonly arose in the middle and lower third of the esophagus. FDG PET-CT was done in 40(71.4%) patients, contrast-enhanced CT instead of PET-CT was done in 14(25%) as there was obvious metastasis in CT scan and 2 (3.6%) patients defaulted before completion of staging work up .However, these 2 defaulters had CT thorax showing metastatic disease. Brain imaging at baseline was available as MRI brain in 14 (25%) cases, FDG PET-CT in 30(53.6%) cases, while in the remaining 12 patients (21.4%) no cross-sectional brain imaging was available. Of the total of 56 patients analyzed, 29 (51.8%) had LD-SCCE, and 27 (48.2%) had ED-SCCE.
Pathology findings:
The tumors showed uniform histology in all the cases, characterized by crushed hyperchromatic cells with scant to absent cytoplasm, prominent nuclear molding and absent to inconspicuous nucleoli. The immunohistochemistry panel utilized for diagnosis included synaptophysin, chromogranin and CD56. All patients had at least two of the above three positive to confirm the neuroendocrine nature of the tumor. The proliferation (Ki-67) index was high (> 70%) in all these tumors (Supplementary Fig. 1).
Treatment (Fig. 1)
LD-SCCE
Of 29 patients, 28 (96.6%) were considered for curative intent therapy while 1 patient (3.4%) was declared best supportive care upfront because of poor Eastern Cooperative Oncology Group (ECOG) performance status (PS).Of 28 patients considered for tumor-directed therapy, local therapy was delivered in 23 (radiotherapy in 19, surgery in 4). The remaining 5 patients who were planned, but did not undergo any local therapy included 3 patients who defaulted post neoadjuvant chemotherapy (NACT) and 2 patients who died from chemotherapy-related toxicities on NACT. NACT and adjuvant chemotherapy were administered in 23(79.3%) and 4(13.8%) cases, respectively. Concurrent chemoradiation (CRT) was planned for all patients undergoing radiation as local therapy, except one patient, who was deemed unfit for chemotherapy. Eleven (37.9%) patients received PCI.
ED-SCCE
Although 24 of the 27 patients (88.9%) with ED-SCCE were advised chemotherapy, 19 (70.4%) received palliative chemotherapy and 7 (25.9%) patients received PCI.
Chemotherapy:
The once-in-3-weeks etoposide and platinum (cisplatin/carboplatin), EP doublet regimen was used in 44 of 46 patients (95.6%) who received chemotherapy. Carboplatin was administered to 28 (60.9%) [13 (28.3%) with LD-SCCE and 15 (32.6%) with ED-SCCE]of 46 patients receiving chemotherapy. The reasons for using carboplatin instead of cisplatin included low glomerular filtration rate (GFR)[9 in LD-SCCE and 5 in ED-SCCE), advanced age and the clinician's judgment regarding chemotherapy tolerance[3 in LD-SCCE and 8 in ED-SCCE), comorbidity [1 in LD-SCCE and 2 in ED-SCCE). One patient received once-a-week paclitaxel and carboplatin as concurrent CRT and one received the same regimen in the palliative setting. The median number of EP cycles in patients with LD-SCCE and ED-SCCE were 4 and 6, respectively.
Response Rate:
Radiologic responses to treatment were available for 15 of 19(78.9%) patients with LD-SCCE undergoing non-surgical modality of therapy and 15 of 19(78.9%) patients with ED-SCCE who received chemotherapy. Responses in patients with LD-SCCE and ED-SCCE included complete response(CR) in 9(9/15, 60%)and 2(2/15,13.3%), partial response in 5(5/15, 33.3%) and 8(8/15, 53.3%), stable disease − 0(0/15) and 3(3/15 ,20%), and progressive disease in 1(1/15, 6.67%) and 2 (2/15, 13.3%), respectively.
Surgery:
Five LD-SCCE patients were planned for surgery. Of these, four underwent esophagectomy with three-field lymph node dissection (cervical, mediastinal, and perigastric lymph nodes), while one patient defaulted post NACT. All 4 patients had R0 resection and had received NACT with EP − 3cycles in two cases, 4 cycles in one case and 6 cycles in one case. Histopathology revealed scanty residual tumor in one case, one focus of in situ carcinoma in one case, while there was evidence of residual invasive carcinoma in the remaining two cases. None of these patients received adjuvant chemotherapy.
Radiotherapy:
The median dose of radiotherapy delivered as external beam radiotherapy (EBRT) in LD-SCCE was 63Gy/35# to loco-regional site. All received conventional fraction EBRT with 45Gy/25# in one, 50.4Gy/28# in two, 55.8Gy/31# in one, 59.4Gy/33# in one, 60Gy/30# in three and 63Gy/35# in the remaining eleven patients. Two patients with ED-SCCE received palliative RT to the local site. The PCI dose regimens included 24.75Gy/11# in six patients, 25Gy/10# and 24Gy/8# in two each with a median dose of 24.75Gy/11# in both LD and ED-SCCE.
Chemotherapy-related toxicities
LD-SCCE: The toxicity data were available for 21 patients. Significant grade ≥ 3 hematological toxicities were:febrile neutropenia (FN) in 7(7/21,33.3%), thrombocytopenia in 2(2/21,9.5%), and anemia in 2(2/21, 9.5%) while the non-hematological toxicities included-chemotherapy induced nausea and vomiting (CINV) and fatigue in 2 each (2/21,9.5%), gastrointestinal (GI) toxicity in 4 (4/21, 19.0%), dyselectrolytemia in 3 (3/21,14.3%), and hepatorenal syndrome in 1 (1/21, 4.8%). Four patients(4/21, 19%) required dose modification or omission during chemotherapy. Deaths attributable to chemotherapy occurred in 2 patients (2/21, 9.5%), one due to febrile neutropenia and another due to hepatorenal syndrome. Overall, 9 patients (42.9%) experienced grade ≥ 3 chemo-toxicity.
ED-SCCE
Toxicity data were available for 17 of 27 patients with ED-SCCE. Notable hematological grade ≥ 3 toxicities were-FN in 4(4/17,23.5%), thrombocytopenia in 1(1/17,5.9%), and anemia in 3(3/17, 17.6%), while the non-hematological toxicities included CINV in 1(1/17,5.9%),fatigue in 3 (3/17, 17.6%),GI toxicity in 1 (1/17, 5.9%),dyselectrolytemia in 2(2/17, 11.8%), and renal toxicity in 1 (1/17, 5.9%). Six (6/17,35.3%) patients required dose modification or omission during chemotherapy. No patient with ED-SCCE died of chemotherapy toxicity. Overall, 10 patients (58.8%) experienced grade ≥ 3 chemotherapy related toxicity.
Survival (Fig. 2)
With a median follow up of 42.2(IQR, 20-61.5) months, the median EFS and OS for the entire cohort were 10.2 (95% CI 7.5–12.9) months and 15.4 (95% CI 12.7–18.2) months, respectively.
LD-SCCE
The median follow-up in surviving patients was 39.2(IQR 27.6–105.6) months.
A)EFS and patterns of failure: At the time of analysis, 22(75.9%) patients experienced events for EFS (recurrence/progression/deaths:17/2/3).Of 19 recurrences/progressions, 2 had both locoregional and distant failures, and 17 had distant-only failures. Only 2 patients failed in the brain. The median EFS was 15.9 (95% CI, 0–30.2)months. The estimated 3-year EFS was 30.5%.
B)OS: There were 19 deaths, including the 2(2/29, 6.9%) chemotherapy-toxic deaths and one (1/29, 3.5%)death due to sepsis post-surgery. The median OS was 22.9 (95% CI 1.8–44.1) months, and the 3-year OS estimate was 41.7%.
ED-SCCE
The median follow-up in surviving patients was 24.4(IQR 8.5–61.5) months.
A) EFS and patterns of failure: At the time of analysis, 25 (92.6%) patients had experienced events for EFS (recurrence:22; progression:2, death due to non-cancerous cause:1). The median EFS was 7.8 (95% CI, 6.6-9.0)months. The projected 3-year EFS was 3.9%.
B) OS: The median OS was 11.8 (95% CI 7.3–16.4) months. The estimated OS at 3years was 3.9%.
Relapse in brain:
Of all the patients with LD-SCCE, 18 (62.1%) patients completed the planned therapy without disease progression and were advised PCI. However, some of these patients underwent PCI and some didn’t. Similarly, PCI was considered for 10 (37%) patients with ED-SCCE who did not have brain metastasis and attained CR or PR at the completion of the first line of chemotherapy. Of these 28 patients, 18 (18/28, 64.3%) received PCI (LD-SCCE:11, ED-SCCE:7). A total of 5 patients relapsed in the brain(LD-SCCE:2, ED-SCCE:3) (Supplementary Fig. 2)
Prognostic factors (Table 2)
Table 2
Significant factors of univariate and multivariate EFS and OS analysis
| | Whole cohort (n = 56) | LD-SCCE (n = 29) | ED-SCCE(n = 27) |
| | [HR=, p=] | [HR, p=] | [HR, p=] |
EFS, univariate analysis |
Age at diagnosis (in years) | ≥ 60 | -- | | -- |
< 60 | -- | HR = 0.2, p = 0.02 | -- |
History of substance abuse | Yes | | | -- |
No | HR = 0.3, p = 0.004 | HR = 0.65,p = 0.03 | -- |
Gender | Female | | -- | -- |
Male | HR = 0.5, p = 0.006 | -- | -- |
LDH | Raised | | -- | -- |
Not raised | HR = 0.3, p = 0.004 | -- | -- |
Intent of treatment | Palliative | | -- | -- |
Curative | HR = 0.3, p < 0.001 | -- | -- |
EFS, multivariate analysis |
History of substance abuse | Yes | | -- | -- |
No | HR = 0.12, p = 0.001 | -- | -- |
OS, univariate analysis |
Age at diagnosis (in years) | ≥ 60 | -- | | |
< 60 | -- | HR = 0.2, p = 0.004 | -- |
ECOG PS | > 1 | | -- | -- |
0–1 | HR = 0.44, p = 0.04 | -- | -- |
Intent of treatment | Palliative | | -- | -- |
Curative | HR = 0.3, p = 0.00 | -- | -- |
LDH | Raised | | -- | -- |
Not raised | HR = 0.32, p = 0.007 | -- | -- |
Site of primary | Cervical, upper 1/3rd | | -- | |
middle-third | | -- | worst prognosis |
lower-third, abdominal | | -- | HR = 0.3,p = 0.002 (best prognosis) |
OS, multivariate analysis |
No significant prognostic factor |
Entire cohort: In univariate analysis, female sex, history of substance abuse, palliative intent treatment and raised LDH predicted for poorer EFS. History of substance abuse was the only independent prognostic factor for EFS while no factor independently prognosticated OS.