Management of Mediastinal Relapse after Treatment with Stereotactic Body Radiotherapy or Accelerated Hypofractionated Radiotherapy for Stage I/II Non–Small-Cell Lung Cancer

Abstract

Purpose/Objective(s)

Regional failures occur in up to 15% of patients treated with stereotactic body radiotherapy (SBRT) for stage I/II lung cancer. This report focuses on the management of the unique scenario of isolated regional failures.

Methods

Patients treated initially with SBRT or accelerated hypo-fractionated radiotherapy were screened for curative intent treatment of isolated mediastinal failures (IMFs). Local control, regional control, progression-free survival, and distant control were estimated from the date of salvage treatment using the Kaplan–Meier method.

Results

Among 160 patients treated from 2002 to 2012, 12 suffered IMF and were amenable to salvage treatment. The median interval between treatments was 16 months (2–57 mo). Median salvage dose was 66 Gy (60–70 Gy). With a median follow-up of 10 months, the median overall survival was 15 months (95% confidence interval, 5.8–37 mo). When estimated from original treatment, the median overall survival was 38 months (95% confidence interval, 17–71 mo). No subsequent regional failures occurred. Distant failure was the predominant mode of relapse following salvage for IMF with a 2-year distant control rate of 38%. At the time of this analysis, three patients have died without recurrence while four are alive and no evidence of disease. High-grade toxicity was uncommon.

Conclusions

To our knowledge, this is first analysis of salvage mediastinal radiation after SBRT or accelerated hypofractionated radiotherapy in lung cancer. Outcomes appear similar to stage III disease at presentation. Distant failures were common, suggesting a role for concurrent or sequential chemotherapy. A standard full course of external beam radiotherapy is advisable in this unique clinical scenario.

Regional relapse after stereotactic body radiotherapy (SBRT) in early-stage lung cancer is seen in up to 10% to 15% of patients.¹ Inoperable patients with metastases to regional lymph nodes (American Joint Committee on Cancer Staging 7th edition, N1 or N2) at presentation are treated with chemoradiotherapy, and it therefore stands to reason patients with regional failures would benefit from a similar treatment.

Despite this logical argument, the successful salvage rate after external beam radiotherapy (EBRT) with or without chemotherapy and the toxicity profile is limited to case reports or extrapolation of radiotherapy salvage after surgical resection and thus poorly understood.^2–4 This report describes our institutional experience and management paradigm for regional failures.

MATERIALS AND METHODS

Data Acquisition

Between 2003 and 2011, 12 patients among 160 treated with SBRT or accelerated hypofractionated radiotherapy (AHRT) for stage I and node negative stage II non–small-cell lung cancer met study criteria. A retrospective chart review was conducted of all patients and 25 suffered recurrence within the hilar or mediastinal lymph nodes. Each patient was restaged with diagnostic computed tomography (CT), and positron emission tomography (PET)/CT imaging and had biopsy-proven disease. Of the 25 patients with regional recurrence, 15 had isolated mediastinal failure (IMF) and 12 were treated with definitive and curative intent. IMF was defined as disease within the ipsilateral hilum and/or sites within the mediastinum in the absence of local recurrence of the originally targeted lung lesion or sites outside of the original lobe treated, including other lobes, contralateral lung, or sites outside the thorax.

Salvage EBRT Treatment

Gross disease only was targeted based on PET/CT and diagnostic CT imaging. No additional expansion was made for a clinical target volume, and no elective nodal irradiation was employed. A standard 0.5 to 1 cm planning target volume was applied. Dosimetric parameters including field arrangement, beam energy, and customized blocking to cover the tumor or spare nearby critical structures were based on the judgment of the treating radiation oncologist. Composite plans were generated of current and prior courses in 10 of the 12 patients. An illustration of an initial, salvage, and composite plan with associated internal target volume and planning target volume (PTV) contours, as well as relevant isodose curves, is shown in Figure 1.

FIGURE 1.

Illustration of initial treatment, salvage treatment, and composite radiation plan. ITV, internal target volume; PTV, planning target volume.

Initial treatments included SBRT (10 patients), AHRT (two patients), and AHRT/SBRT (one patient with bilateral stage I cancers). AHRT was selected as the treatment choice based on the location and size of the lesion at the discretion of the treating physician. Median time interval between initial and salvage course was 15 months with a range of 2 to 57 months. The median salvage EBRT dose was 66 Gy and 33 fractions with concurrent or sequential chemotherapy delivered in two and one patient, respectively. The average composite maximum dose combining the initial and retreatment plan was 90 Gy (range, 73–145 Gy). The mean composite lung V20 was 25% (range, 13%–29%) and mean esophageal dose was 10.7 Gy (range, 3.4–33 Gy).

Follow-Up and Statistical Analysis

Patients were considered to have failed locally after evidence of increased size of enhancing tumor in the originally treated region or within the same lobe of the lung. Patients were screened for regional failures defined as new nodal disease in the hilum or mediastinum, including the area targeted at time of salvage EBRT. Distant failures were defined as disease in the contralateral lung or in sites outside of the thorax. PET scans and/ or biopsy was employed to assist with differentiating radiation related lung changes with local or regional recurrence. Toxicity was gathered retrospectively and scored per the Common Terminology Criteria for Adverse Events (CTCAE v4.0).

Time to event data was recorded from date of salvage EBRT completion for all measures except local control (LC), where the date of original SBRT or AHRT completion was used. Kaplan–Meier estimates were completed for progression-free survival (PFS), overall survival (OS), LC, regional control, and distant control (DC)..

RESULTS

Patients

Table 1 outlines all 12 patients treated for IMF. Ten were treated with stage IA disease initially, whereas two patients were treated for T3N0 tumors, both with two separate lesions in the same lobe. Average tumor size for the initial treatment was 1.5 cm with a range of 0.7 to 2.8 cm. The locations of the regional failures were as follows: isolated hilar failure in three patients, single ipsilateral mediastinal station in four patients, hilar and single mediastinal station in two patients, and multiple mediastinal stations in three patients. The median age at salvage treatment was 66.

TABLE 1.

Patient Characteristics and Outcomes

	Sex/Age at Salvage	PS	Stage	Pathology	Initial Course	Interval (mo)	Salvage Course	CTX	Local Failure/Time to Failure (mo)	Regional Failurea/Time to Failure (mo)	Distant Failure/Time to Failure (mo)	Surviving/Cancer-Related Death	NED
					Dose (Gy)/Fractions		Dose (Gy)/Fractions
1	Female/79	1	T1N0b	Adenocarcinoma	54/3, 70.2/26b	2	60/30	No				No/No	Yes
2	Male/71	1	T3N0c	SCC	50/5	5	62/23	No			Yes/7	No/Yes	No
3	Male/55	1	T1N0	Adenocarcinoma	54/3	7	66/33	Yes				Yes	Yes
4	Female/65	1	T1N0	Adenocarcinoma	50/5	12	65/26	No			Yes/6	Yes	No
5	Female/53	0	T3N0	SCC	70.2/26	15	60/30	No				Yes	Yes
6	Male/67	0	T1N0	Adenocarcinoma	70.2/26	15	66/33	No				No/No	Yes
7	Female/74	2	T1N0c	Adenocarcinoma	50/5	16	66/33	No			Yes/6	No/Yes	No
8	Female/59	2	T1N0	Adenocarcinoma	45/5	17	70.2/26	No				No/No	Yes
9	Female/61	1	T1N0c	Adenocarcinoma	54/3	19	65/36	Yes			Yes/14	No/Yes	No
10	Female/60	1	T1N0	Adenocarcinoma	50/5	21	66/33	No				Yes	Yes
11	Male/69	1	T1N0	SCC	60/3	51	70/35	No				Yes	Yes
12	Female/85	0	T1N0	Adenocarcinoma	54/3	57	70.2/26	No	Yes/66		Yes/5	No/Yes	No

^aNo patient suffered subsequent regional failure after salvage treatment.

^bPatient received SBRT(54 Gy in three fractions) and accelerated hypofractionation (70.2 Gy in 26 fractions) for contralateral stage I lung cancers.

^cTwo separate stage I lung primaries treated with SBRT: bilateral or tumors or tumors in separate lobes of the same lung.

PS, performance status; CTX, concurrent chemotherapy; NED, no evidence of disease; SCC, squamous cell carcinoma; SBRT, stereotactic body radiotherapy.

LC, Regional Control, DC, PFS, and OS

Median follow-up was 10 months with a range of 2 to 49 months. A single patient failed within the same lobe treated 66 months after the initial SBRT course yet after the salvage radiation course and was scored as local relapse while also suffering synchronous brain metastases. No regional failures occurred after the salvage radiation course. The local and regional failure-free survival at 2 and 5 years was 100% and 92%, respectively.

Distant relapses were noted in five of the 12 patients (42%). The estimated 1- and 2-year DC was 56% and 38%, respectively. The 2-year PFS rate was identical to the DC rate with a median PFS of 14 months (95% confidence interval [CI], 5–49 mo). At the time of this analysis, four patients died following distant relapse, three have died without recurrence, one is alive with documented distant relapse, and four are alive with no sign of disease (Fig. 2). The estimated 1- and 2-year OS was 58% and 29%, respectively, with a median of 15 months (95% CI, 6–37 mo). The median OS from original treatment was 38 months (95% CI, 17–71 mo).

FIGURE 2.

Kaplan–Meier estimates for distant control (A), PFS (B) from time of isolated mediastinal failure, and overall survival from the time of isolated mediastinal failure and from initial treatment (C). All progression events within 3 years were distant; therefore, the distant control and PFS curves are identical. PFS, progression-free survival.

Toxicity

Four of the 12 patients (33%) experienced grade 2 acute or late toxicity, whereas one patient suffered late grade 3 toxicity from the salvage course of radiation. Acute grade 2 toxicity was seen in four patients mainly related to esophagitis. Two patients experienced late toxicity with one requiring supplemental oxygen (grade 3) and one with sustained dysphagia (grade 2) (Table 2).

TABLE 2.

Clinical Outcomes

	Cumulative Incidence (n = 12)	2-Year Failure Free Survival (n = 12)
Outcome	n (%)	% (95% CI)
Any failure	6 (50)	38% (7%–70%)
Local	1 (8)a	100%a
Regional	0	100%
Distant	5 (42)	38% (7%–70%)
Overall survival		29% (5%–61%)
Toxicityb	Acute	Late
Esophagitis	3 (25)	1 (8)
Dyspnea	1 (8)	1 (8)
Any	4 (33)	2 (17)

^aSingle failure occurred at 66 mo.

^bAll toxicities were grade 2 except a single patient with grade 3 late dyspnea requiring supplemental oxygen.

n, number; CI, confidence interval.

DISCUSSION

This brief report focuses on the management of regional recurrences in early-stage lung cancer. Although regional failures occur in up to 15% of these patients,^5–7 a minority will be candidates for further thoracic radiotherapy. Many will have synchronous local or distant failure or a performance status that precludes additional treatment.⁵ The crude IMF rate in our experience was 9%. Unfortunately, chemotherapy is unable to provide durable tumor control and has modest response rates at the cost of substantial toxicity.⁶ Salvage radiotherapy for recurrence after surgical resection has been described. Kelsey et al.⁴ reported on 25 patients with stage I disease treated for recurrence after surgery with radiotherapy or chemoradiotherapy more than a 12-year period. Median survival was 17 months, and the predominate first site of failure was distant. Their report is consistent with earlier reports for radiotherapy salvage after surgery and with our current series.^3,4,7 Therefore, given the near certainty of progression with chemotherapy treatment and the postoperative radiotherapy salvage results, we routinely evaluate and treat patients with IMF with radiotherapy as there is some hope for salvage.⁶

Interestingly, although not unexpectedly, outcomes for stage I/II patients at the time of IMF seem to mirror stage III disease measured from initial treatment. The median OS in cooperative group trials with locally advanced lung cancer has ranged from 12 to 24 months with distant relapse rates up to 40%.⁸ Among our small cohort, median OS was 15 months, and five of the 12 patients (42%) experienced subsequent distant relapse. With the predominant pattern of failure being distant, this would argue for chemotherapy in this scenario either concurrently or sequentially.

Although IMF is uncommon, providers are reluctant to consider retreatment out of concern for toxicity in part based on data for retreatment at the same site in the thorax.⁹ Multiple reports have detailed reirradiation focusing mainly on local failure as opposed to regional failure with either salvage surgery,^10,11 radiation,^12–14 or even SBRT.¹⁵ Nevertheless, although reirradiation at the same location carries substantial risk, the mediastinum has generally not received a prohibitive dose after SBRT or AHRT as demonstrated in our composite dosimetry. The low V20 and mean esophageal rates reflect the small fields used in SBRT and the avoidance of dose spillage from the initial plan into the mediastinum. Furthermore, our low acute toxicity rate of 33% grade 2 and 0% grade 3 highlights the separation geographically of the treatments and thus its safety. We suggest targeting gross disease only at salvage omitting elective coverage of the hilum or adjacent areas in the mediastinum. This decision appears well justified when viewed in context with the data demonstrating safe omission of elective nodal irradiation in stage III patients, the present study’s low toxicity rates paired with our excellent control rates.

CONCLUSIONS

Although IMF after SBRT or AHRT is uncommon, patients can be salvaged safely and effectively with fractionated EBRT. Although a small sample size hinders our ability to draw definitive conclusions, making the parallel to stage III lung cancer, concurrent chemotherapy should be considered given the high distant failure rate if the patient’s performance status permits.