Comparison of outcomes for patients with medically inoperable Stage I non–small-cell lung cancer treated with two-dimensional vs. three-dimensional radiotherapy

Introduction

The first and only randomized study comparing surgery and radiotherapy as treatment for operable non–small-cell lung cancer (NSCLC) was published in 1963 and showed surgery to be superior to radiotherapy alone. Since then, surgery has been the established treatment of choice for Stage I NSCLC. However, there is a distinct group of patients who are considered medically inoperable because of poor cardiopulmonary function or other comorbidities. In addition, some patients refuse surgery. In these patients, radiotherapy alone has been shown to provide an acceptable alternative. Series on radiotherapy alone as treatment for Stage I NSCLC report 5-year survival rates ranging from 5% to 30%  compared with surgery, which confers 5-year survival rates of 50–67%  Inadequate radiation dose has been suggested as one of the factors that lead to inferior outcomes seen in patients treated with radiotherapy alone. Historically, it has been difficult to increase the dose to lung tumors without exceeding the tolerance of normal lung, which is of particular importance in medically inoperable patients who largely have low pulmonary reserves. With the advent of three-dimensional (3D) conformal radiotherapy and smaller margins, higher doses have been achieved without sacrificing surrounding tissue. However, highly conformal planning raises the concern of whether the entire tumor volume is treated, particularly when accounting for tumor motion through the respiratory cycle. Since 1997, University of Texas M.D. Anderson Cancer Center has adopted 3D conformal treatment planning. This retrospective analysis was performed to assess the difference in outcomes between reduced field radiation with 3D conformal radiotherapy and historical two-dimensional (2D) planning.

Discussion

Patients who received 3D conformal radiotherapy in this series achieved a 5-year OS rate of 36% and a DSS rate of 68%. These rates compare favorably with those found in the literature, which show 5-year OS rates in the range of 5–30% and DSS rates between 13% and 39% for Stage I NSCLC treated with radiotherapy alone. Survival rates of patients who received 2D radiotherapy in this study are within those ranges with a 5-year OS rate of 10% and DSS rate of 29%. With the exception of two studies, previous series did not uniformly treat patients with 3D conformal radiotherapy, nor did they stratify patients to treatment technique and perform subset analyses.

The results of this study indicate that 3D conformal radiotherapy does not compromise the outcomes of patients with Stage I NSCLC when compared with 2D radiotherapy. In fact, survival rates and LRC are superior in patients treated with 3D conformal radiotherapy. One logical reason for this is higher radiation dose with 3D treatment; however, there was no significant dose difference seen in this study. The toxicity of larger fields used in 2D radiotherapy likely had a negative impact on survival. In addition, elective nodal irradiation that was received by 41% of patients treated with 2D radiotherapy further contributed to toxicity. This is supported by the association of mediastinal irradiation with decreased OS on univariate analysis (hazards ratio 1.64, p = 0.01). With the exception of 3 patients, the mediastinum was not included in the treatment field for 3D patients. However, we assume that some patients received incidental dose to the mediastinum

In the past, elective nodal irradiation was a controversial topic, but since then elective nodal irradiation has not been shown to improve OS, DSS, or LRC . The incidence of regional relapse with the omission of elective nodal irradiation  is quite low at 0–3.2% therefore, elective nodal irradiation is no longer recommended.

A difference in outcomes between 2D and 3D treatment in patients with T1 tumors was shown in this study; however, there was no difference seen with T2 tumors. A possible explanation is the inability of radiotherapy alone to control T2 lesions, resulting in a small fraction of patients achieving local control or being alive at 5 years. Numerous Stage I NSCLC studies have identified T stage and tumor size as poor prognostic factors. Reported 5-year OS rates for T2 lesions from the literature are significantly lower than T1 lesions and range between 4% and 24% . According to the teachings of Fletcher, doses up to 100 Gy may be required to sterilize 3-cm lung tumors, which may explain the inferior outcomes of radiotherapy alone with T2 tumors. Another factor that may have led to the lack of difference between 2D and 3D in T2 tumors is the relatively smaller differences in field sizes between the two techniques for larger lesions because radiation oncologists may have reservations in treating with the full standard margins, particularly in medically inoperable patients. As a result, the difference in outcomes that resulted from the toxicity with larger fields is minimized when comparing 2D with 3D radiotherapy.

Despite improvements in outcomes with modern advances in radiotherapy, survival rates still fall far behind surgical series. There are a number of factors that contribute to the inferior outcomes seen in radiotherapy series. First, patients treated with radiotherapy alone are clinically staged. It has been shown that 11–23%  of patients with Stage I NSCLC and radiographically negative lymph nodes have occult positive nodes after pathologic staging. Another factor is the difference in patient population who undergo surgery vs. radiotherapy. Patients who are referred for radiotherapy are older, have lower performance status, and are medically inoperable due to the presence of comorbidities. Many patients who are referred for radiotherapy have severe pulmonary compromise with forced expiratory volumes less than 50%, which generally deems a patient medically inoperable. The 5-year survival in this group of patients even in the absence of lung cancer is 41%. The observation that definitive radiotherapy studies with the greatest number of patients who refused surgery showed improved survival from the absence of competing morbidities demonstrates the impact of comorbidities on survival rates.

Although there are multiple factors that confound the results of radiotherapy series, inadequate dose has been identified as contributing to inferior outcomes when compared with surgery. The impact of radiation dose has been addressed by numerous investigators. Cooper et al. first found improved survival with higher doses of radiation for NSCLC . Since then, several studies have had similar findings, A Radiation Therapy Oncology Group, trial showed that local control is dependent on tumor dose, with intrathoracic failure rates of 48% with 40 Gy, 38% with 50 Gy, and 27% with 60 Gy. Dosoretz et al. reported a 22% DFS at 24 months for patients who received 50–60 Gy, 33% for doses of 60–70 Gy, and 50% for patients who received a dose of 70 Gy or more. Three-dimensional conformal radiotherapy provides an opportunity to increase total dose to the tumor while lowering dose to normal lung and surrounding critical structures. To this end, RTOG 93-11 was designed as a dose-escalation trial to evaluate toxicity using 3D conformal radiotherapy in patients with inoperable NSCLC. Dose was safely escalated to 83.8 Gy in patients who received greater than 20 Gy to less than 25% of total lung volume and 77.4 Gy in patients who received greater than 20 Gy to 25–36% of total lung volume. Dose of 90.3 Gy resulted in two dose-related deaths. This trial approached 100 Gy, but was unable to reach this dose without significant toxicity.

To reduce toxicity as well as improve tumor dose, strategies have been developed to account for respiration-induced organ motion. One such strategy is deep inspiration breath hold (DIBH). Hanley et al. looked at lung motion fluoroscopically and found a 3% mean lung area change for DIBH and an 11% change with normal breathing. Several other studies had consistent findings, so it was suggested that PTV can be reduced when using DIBH. In one study, average prescription dose was increased from 69.4 Gy to 87.9 Gy with DIBH. Respiratory-gated treatment is another strategy that is being investigated. A safe reduction in margins has been achieved by investigators. Respiratory-gated treatment may be more tolerable for lung cancer patients, particularly those who are medically inoperable.

Radiation techniques other than 3D conformal radiotherapy have been investigated in hopes of achieving higher doses while minimizing toxicity. Stereotactic fractionated radiotherapy has been studied and initial results are promising. Onishi et al. reported 3-year and 5-year OS rates of 56% and 47%, respectively, for 245 patients with Stage I NSCLC treated with stereotactic hypofractionated irradiation. The local control rate was 91.9% with biologic equivalent doses ≥100 Gy. Five-year OS rates in operable patients treated with biologic equivalent doses ≥100 Gy were 90% for T1 tumors and 84% for T2 tumors. These rates surpass those of surgical series and thus are quite encouraging. However, long-term toxicities of hypofractionated radiotherapy for NSCLC remain unknown.

There are limitations to this study, including those that are inherent to retrospective studies, such as unknown selection biases. In addition, this series had a greater percentage of patients with T1 tumors, which likely increased survival rates. Another potential confounder is a relatively higher proportion of patients who were medically operable but refused surgery. Last, patients in the 2D group were treated in an earlier time frame when radiographic imaging was less accurate, and thus some patients may have been understaged. With modern improvements and the advent of PET scans, patients treated with 3D radiotherapy were more likely to be accurately staged, because 57 patients treated with 3D technique received a PET scan as part of their initial staging. Several studies have found PET to be more accurate than CT in the staging of lung cancer. Bradley et al. and Hicks et al. reported that 47–50% of patients with Stage I NSCLC were upstaged with PET imaging; therefore, stage migration likely affected the outcomes of this study. However, if this played a major role,one would expect a statistically significant difference in DMFS between the 2D and 3D groups, which was not seen.

Conclusions

This study demonstrates that 3D conformal radiotherapy is not only equivalent to 2D radiotherapy, but in fact may significantly improve survival in patients with Stage I NSCLC treated with radiotherapy alone. Given that there was no significant difference in delivered dose, the inferior outcomes with 2D radiotherapy were likely from toxicity from larger field sizes and damage to normal tissue that was particularly critical to medically inoperable patients. For patients who are medically inoperable or refuse surgery, 3D conformal radiotherapy is an acceptable treatment modality. As lung cancer screening trials are under way, there will be an anticipated increase in the incidence of Stage I NSCLC, so it is imperative that investigation of optimal treatment for this disease be assertively continued.