Discussion 

Surgical treatment of Stage I NSCLC with lobectomy and hilar/mediastinal lymph node removal has been the standard against which alternative treatments have been compared. For patients unable to tolerate this procedure because of medical comorbidities such as chronic obstructive pulmonary disease, cardiac disease, diabetes mellitus, and vascular disease, options include a less extensive surgery such as wedge resection, local radiotherapy, or observation. However, studies have shown that patients managed by observation alone have 5-year survival rates of less than 10%, and that even with severe competing medical problems the majority of untreated patients die of their cancer. Wedge resection and conventionally fractionated radiotherapy have been associated with high rates of local tumor recurrence. Because of these factors, Phase I and II trials were completed at Indiana University to investigate stereotactic body radiation therapy (SBRT) using a hypofractionated regimen with large biologically effective doses (BEDs).

Our Phase I study showed encouraging results for a high potency SBRT regimen of three high-dose fractions administered to a frail population of medically inoperable patients. This trial was a dose escalation study, stratified by tumor size, starting at a dose of 8 Gy × three fractions (24 Gy) and escalating by 2 Gy/fraction in cohorts of 3 patients each. The maximum tolerated dose (MTD) was not reached for T1 tumors and T2 tumors less than 5 cm even up to dose levels of 20 to 22 Gy × three fractions (60–66 Gy). For patients with tumors measuring 5 to 7 cm, the dose was escalated to 24 Gy × three fractions (72 Gy), which was considered dose limiting. Patients treated at 22 Gy/fraction have not had late toxicity after follow-up of more than 2 years. Of 47 patients, 10 developed local failure. Nine of the local failures occurred at doses of 16 Gy × three fractions or less.

No attempt was made to exclude from our Phase II study patients with centrally located tumors. In the preliminary report of our study, we demonstrated a significantly decreased time to Grade 3 to 5 toxicity in patients with centrally located tumors. However, our mature results show a relatively low risk of toxicity for the complete study population. In addition, after tumor location was re-classified based on the criteria in RTOG 0236, our current results show that the difference in toxicity between central and peripheral tumors does not reach statistical significance despite the fact that the estimated toxicity rates in the patients with central tumors (6 patients with high-grade toxicities out of 22; 27.3%) were almost three times the rate in the group of patients with peripheral tumors (5 patients with severe toxicities out of 48; 10.4%). This is most likely caused by the small number of severe toxicities observed in the study. The safety of this regimen was also demonstrated in a separate analysis of our serial pulmonary function test (PFT) data. We found no significant difference in survival or post-treatment PFT values between patients in the lowest quartile of pretreatment forced expiratory volume in 1 s (FEV1 < 0.7 L) and diffusion capacity for carbon monoxide (DLCO <8.4 ml/min/mm Hg) and those patients in the upper three quartiles. We did find a statistically significant decline in DLCO over time, which is consistent with our current finding of oxygen dependence over time after treatment. However, it must be noted that declines in FEV1 and DLCO are to be expected over time, especially in a population that generally had a significant smoking history.

The promising results of our institutional trial, as well as increasing international publications on SBRT, resulted in a national Phase II clinical trial, RTOG 0236. Based on our early concerns about toxicity in patients with centrally located tumors, RTOG 0236 excluded patients with PTV volumes encroaching on a 2-cm margin around the mediastinum and major airways. This trial involved treatment of biopsy-proven Stage I NSCLC up to 5 cm, to doses of 20 Gy × three fractions = 60 Gy, at institutions undergoing an extensive accreditation process. The trial has completed accrual and results are pending.

A similar multicenter prospective trial is ongoing in Japan, treating both operable and inoperable patients with doses of 12 Gy × four fractions = 48 Gy. This is based on institutional data from Kyoto University Hospital, where patients with Stage I NSCLC at all locations were treated with this regimen, without the finding of increased toxicity in patients with centrally located tumors. However, the locations of tumors were not specified. A Phase I trial specifically addressing centrally located tumors is under development by the RTOG.

Despite the observed toxicity in some patients, the results of our Phase II protocol are encouraging. Patients with peripherally located tumors had relatively low treatment toxicity. In addition, with median follow-up of 50.2 months, only 5.7% of patients experienced local recurrence, for a Kaplan-Meier 3-year local control rate of 88.1%, which rivals the control rates reported after lobectomy. As expected, this frail group of patients had a significant rate of intercurrent death, which influenced OS.

Other published series of SBRT have similarly shown high rates of local control when high BEDs are used. One of the largest is a retrospective study pooling data from multiple Japanese institutions. This included 245 patients with early Stage NSCLC (T1, n = 155; T2, n = 90) treated with a variety of SBRT regimens. The median calculated BED was 108 Gy (range, 57-180 Gy). With a median follow-up of 24 months, the local progression rate was 14.5% (8.1% for BED ≥ 100 Gy vs. 26.4% for BED < 100 Gy). The 3-year OS rates for BED ≥ 100 Gy and BED < 100 Gy in medically operable patients were 88.4% and 69.4%, respectively. In a pooled analysis of patients treated with SBRT in Sweden and Denmark, 138 patients with medically inoperable NSCLC were prescribed doses of 30 to 48 Gy in two to four fractions (most commonly, 15 Gy × three fractions prescribed to 65% isodose). With a median follow-up of 33 months, the local control rate was 88%. Distant metastasis occurred in 25% of the patients, and the 3- and 5-year OS rates were 52% and 26%, respectively. In addition, the results of a prospective Phase I/II trial of SBRT have been reported from Japan. This trial used a dose regimen of 12 Gy × four fractions (total dose of 48 Gy, prescribed to isocenter). A total of 45 patients (32 Stage IA and 13 Stage IB with tumor size <4 cm) were enrolled in the trial, including both medically operable and inoperable patients. With median follow-up of 30 months, the crude local control rate was 98%. Distant metastases occurred in 5 and 4 patients with T1 and T2 tumors, respectively. For Stage IA lung cancer, OS rates after 1 year and 3 years were 92% and 83%, respectively, whereas for Stage IB lung cancer, OS rates were 82% and 72%, respectively.

We now report excellent 3-year local control and survival rates post-SBRT. Mature data, both from the United States and internationally, support the currently open RTOG 0618 study, which prospectively offers select healthy patients with Stage I/IIA NSCLC treatment with SBRT and reserves surgical salvage for the relatively rare local failure.