Analysis of a Novel Protocol of Combined Induction Chemotherapy and Concurrent Chemoradiation in Unresected Non–Small-cell Lung Cancer: A Ten-year Experience with Vinblastine, Cisplatin, and Radiation Therapy

Eugenie Waters; Brian Dingle; George Rodrigues; Mark Vincent; Robert Ash; Rashid Dar; Richard Inculet; Walter Kocha; Richard Malthaner; Michael Sanatani; Larry Stitt; Brian Yaremko; Jawaid Younus; Edward Yu

Disclosures

Clin Lung Cancer. 2010;11(4):243-250. 

In This Article

Results

Clinical and Pathologic Data

The characteristics of the study population are listed in Table 2. There were a total of 294 patients who had a clinical stage of IIIA (138 patients) or IIIB (156 patients) NSCLC and who underwent VCRT. At the time of analysis, 214 (72%) of these patients had died, and the follow-up of the survivors ranged from 1 month to 11 years, with a median of 34 months.

The histology of NSCLC represented in this study was evenly divided, with adenocarcinoma (35%), squamous cell carcinoma (36%) and "other" (29%), the latter primarily consisting of NSCLC unspecified. At diagnosis, all VCRT patients underwent CT scans of the thorax and upper abdomen (100%), a large majority had a bone scan (87%), and more than half had a CT scan of the head (53%). Only 11% of VCRT patients had a PET scan, but over one-third underwent mediastinoscopy (35%). All 294 VCRT-designated patients actually started VCRT chemotherapy (100%), and the large majority also initiated radiation therapy (88%). A subset of 26 patients (9%) had surgery within 6 months following completion of concurrent chemotherapy and radiation therapy (Table 3).

Overall Survival and Comparison with Published Studies

Overall survival for all stage IIIA and stage IIIB NSCLC patients who underwent VCRT was calculated with a Kaplan-Meier survival curve as shown in Figure 1. The MST was 18.2 months and the 5-year OS was 19.8%. The 2, 3, 4, and 5-year survival estimates are listed in Table 4 .5,9–29] For comparison, relevant clinical outcomes from various other clinical trials employing sequential or concurrent platinum-based chemotherapy are presented in the same table. Where these studies had 2 or 3 arms, and differences between these arms is apparent, even if not statistically significant, the statistics displayed are those of the best arm.

Figure 1.

Overall Survival of All Patients With IIIA/IIIB NSCLC Treated With VCRT at the LRCP Between 1996 and 2006 (MST = 18.2 Months)

Overall Survival and Comparison within VCRT Datasets

Comparisons within our own VCRT dataset are presented in Table 5. A variety of metrics, at both 2 and 5 years, along with corresponding calculated median values are displayed. There was no significant difference in OS between those patients whose corrected clinical stage was IIIA NSCLC (n = 138) versus IIIB NSCLC (n = 156; P = .27).

Considering toxicity, 25% of patients experienced at least 1 event that fit the criteria for a CTCAE grade 4 toxicity. Despite this rather large number, there was no significant difference in survival between those who experienced a grade 4 toxicity and those who did not (P = .26). Specific observed toxicities are presented in Table 6. A significant proportion of patients experienced acute hematologic toxicities of any grade during treatment such as neutropenia (68%), thrombocytopenia (45%), and anemia (84%). Similarly, the large majority of patients reported esophagitis (61%) and nausea/vomiting (76%) during treatment. Neuropathy, predominantly grade 2, was the most common late toxicity experienced by patients (44%). Severe (grade 3–4) pneumonitis and esophagitis were only noted in 1% and 3% of patients, respectively.

Given this study's retrospective nature, the information that could be gathered regarding treatment-related deaths with VCRT was limited. Thirty patients (10%) died within the first 4 months of starting treatment. Of these 30 patients who died, 3 died of treatment-related toxicity, at least 8 died of progressive disease, and 14 did not receive radiation.

There was no significant difference in survival outcomes between the patients who received VCRT during the first 5 years of the decade-long study period (1996–2006) versus those who were treated in the second 5-year period (log-rank P = .13). Only 28 of 294 patients who had VCRT changed chemotherapy treatment from cisplatin to carboplatin during the protocol, based on a clinical judgment to reduce the toxicities experienced to that point. Tumor size was documented for all patients. The log-rank test comparing survival for tumor size > 5 cm versus a size ≤ 5 cm was not significant but suggested a trend toward improved survival with smaller tumor size (P = .18). Cox proportional hazard regression indicated a similar trend (P = .06; H, 1.055; 95% CI of HR, 0.498–1.116).

Although 60 Gy was the standard radiation dose for the VCRT protocol, 12% of patients did not receive any radiation therapy. Log-rank tests comparing those who received a nonzero dose of radiation therapy compared with those who received no radiation therapy demonstrated a survival benefit associated with the inclusion of radiation therapy (P = .0001). Univariate Cox regression showed that receiving no radiation therapy was associated with reduced survival (P = .0001; HR, 2.259; 95% CI of HR, 1.512–3.373). Similarly, receiving any amount of radiation totaling less than the standard 60 Gy was also related to reduced survival (P = .0014; HR, 1.632; 95% CI of HR, 1.209–2.202). Figure 2 illustrates the significantly different Kaplan-Meier curves of these 3 patient groups.

Figure 2.

Comparison of Overall Survival of IIIA/IIIB NSCLC Patients Treated at the LRCP Who Received No Radiation Therapy (RT), < 60 Gy RT, and the Full VCRT Dose of RT (60 Gy)

The value of having a combined total dose of at least 500 mg of cisplatin was confirmed with the log-rank test (P = .0032). Sixty-six percent of VCRT patients had a minimum cumulative dose of 500 mg of cisplatin. The Cox proportional hazard regression model demonstrated that both cisplatin dose (P = .0002, HR = 0.999, 95% CI of HR (0.998–0.999)) and vinblastine dose (P = .0001; HR = 0.971; 95% CI of HR, 0.957–0.986) are associated with improved survival time in VCRT patients with IIIA/IIIB NSCLC.

There was a statistically significant survival benefit for the small subset (26 patients) who underwent completion surgery within 6 months following VCRT (log-rank P = .0002, Cox P = .0003). The MST was 31.9 months for this group compared with 18.2 months for all IIIA and IIIB NSCLC patients who underwent VCRT. A summary of the surgery performed is presented in Table 3. All completion surgeries (surgeries within 6 months of completion of VCRT) had clear resection margins except for 1 right-sided pneumonectomy and 1 left-sided lobectomy, which had microscopic involvement at the margins, and 2 surgeries in which resection margins were not documented. In summary, an R0 resection was performed in 22 patients, and a total of 17 patients had their cancer pathologically down-staged, of which 9 were confirmed to have pathologic complete remissions.

In the first VCRT multivariable model 5 variables were most significantly related to survival: completion surgery, radiation therapy (3 dose categories: no RT, < 60 Gy, 60 Gy), cisplatin dose, vinblastine dose, and tumor size. A second VCRT multivariable model included the same variables allowed to enter stepwise. The multivariable analyses supported the assertion that tumor size, radiation dose, and completion surgery are significant in a model of therapy for clinically unresectable NSCLC.

processing....