In the early 1980s, neoadjuvant therapy, a term that implied treatment given before definitive surgical management, began to be used in the treatment of NSCLC. Previously, patients with bulky tumor or extensive nodal involvement usually were not considered surgical candidates and were treated with radiation therapy as definitive treatment.

With the introduction of the concept of neoadjuvant therapy, treatment with chemotherapy, radiation therapy, or both began to be given to reduce bulky tumor or sterilize lymph nodes in an attempt to convert marginally unresectable or unresectable disease to resectable disease. Preoperative or induction therapy theoretically may have advantages in being used before resistant clones of cells have the opportunity to develop.⁷⁹ Directed cell killing has an impact on stimulating the immune sytem before surgery. Treatment before resection allows the clinician to assess tumor responsiveness, helping to identify patients who may benefit from continuation of postoperative adjuvant chemotherapy. Prior to surgery, patients may be better able to tolerate full dosages of neoadjuvant treatment. Finally, neoadjuvant therapy in concept might lessen the incidence of later distant disease when given at a time when only micrometastatic disease may be present.

A number of contrary positions exist regarding neoadjuvant therapy. Tumors might progress during the delay to the time of operation and a previously resectable tumor might become unresectable. Preoperative therapy is associated with systemic toxicity that might cause difficulty during the postoperative recuperation and wound healing. Several chemotherapeutic agents are associated with pulmonary toxicity and can result in reduced pulmonary function. Publications by Roberts,⁶⁸ Martin,⁴⁴ and Novoa⁵² and their colleagues as well as by Torre and Sierra,⁹³ addressing the question of whether patients who have undergone neoadjuvant therapy are at increased risk of postoperative complications, have reported conflicting results. However, there is little doubt that following right pneumonectomy, these patients are at high risk; therefore every effort should be made to avoid right pneumonectomy in these patients.

Resections after preoperative therapy can be extremely difficult and potentially hazardous because of the fibrosis that often results as a response to the therapy. Tissue planes may be obliterated because of mediastinal fibrosis. This is especially significant when there has been a response in the lymph nodes, because the nodes are intimately associated with the pulmonary artery and its branches, often making resection quite challenging. Because of this progressive fibrosis, patients are ideally taken to the operating room 3 to 5 weeks after the completion of neoadjuvant therapy. It is particularly important to have proximal control of the pulmonary artery before undertaking a resection in a patient who has received preoperative therapy. Resections of this type ideally should be undertaken by a surgeon experienced in dealing with complex resections so as to avoid an unnecessary pneumonectomy.

Despite the possible difficulties associated with neoadjuvant treatment, there is reasonably strong evidence from small, randomized trials that neoadjuvant therapy results in improved survival when compared with surgery alone for stage IIIA (N2) disease. However, this has never been confirmed in the type of larger, multi-institutional phase III study that is likely to
convince most in the field. Further, it is difficult to assess the various nonrandomized reports of neoadjuvant therapy. Staging criteria have differed widely, as have inclusion criteria. The criteria used to determine mediastinal lymph node involvement have varied from simple radiologic evidence to pathologic confirmation. The majority of studies do report a 50% to 60% objective response rate, although the complete response rate tends to be less than 15%. The overall response rate is significantly higher than when these same agents are used in patients with disseminated disease. The vast majority of responders to neoadjuvant chemotherapy and approximately half the patients overall are able to go on to surgical resection.^10,37,64,79 Median survival is approximately 18 months for patients receiving combined-modality treatment, including induction chemotherapy (ranging from 8 to 30 months), with 2- to 3-year survival ranging from 25% to 30%. Although it may seem that the median survival of approximately 18 months is superior to that of surgery or radiation therapy alone, one must remember that the patients in these trials represent a select subgroup of stage III patients overall, usually those with the best performance status, and that this survival advantage may reflect patient selection rather than actual improvement with therapy.

Stages I and II NSCLC have traditionally been managed by operation alone. However, the 5-year survival rate for clinical stages I and II ranges from only 50% to 80%. No study to date has conclusively demonstrated any survival advantage of preoperative chemotherapy or radiation for early-stage disease.

Patients who have not received preoperative therapy and are discovered to have mediastinal lymph node involvement (N2 disease) at the time of thoracotomy should likely receive postoperative radiation or chemotherapy or both. Although even in this advanced stage no randomized study has documented improved survival with postoperative adjuvant treatments added to surgery alone, there is sufficient randomized and nonrandomized data that lead many physicians to strongly weigh the benefits of these therapies against the risks of these modalities.

The former international staging system created by a collaboration of the American Joint Committee on Cancer Staging and the International Union Against Cancer (1986) divided stage III lung cancer into two distinct groups.⁴⁹ Stage IIIA comprised those lesions that are potentially resectable and included any N2 disease, any T3 primary tumors, or both. Stage IIIB designated those patients with disease that involves structures that usually preclude resection by a unilateral approach. These included T4 primary tumors and N3 nodal disease. The international staging system published in 1997 by Mountain has changed the categories in IIIA disease. However, most of the studies quoted in this chapter have used the former international staging system, and thus this system, albeit outdated, is used for the most part throughout this chapter.

Approximately 20% of patients with NSCLC present with stage IIIA disease at diagnosis, and another 20% are stage IIIB at presentation.¹⁷ The overall median and 5-year survival for stage IIIA disease are 12 months and 15% to 20%, respectively, and for stage IIIB, 8 months and 0% to 5%.^78,103 Results this poor with disease that is at least potentially resectable have made patients in stage IIIA (in particular N2) a focus for multimodality therapies.

Until recently, patients who presented with even ipsilateral involvement of mediastinal lymph nodes (N2) were considered inoperable and tended to be treated with radiation or chemotherapy alone. This therapeutic nihilism evolved because of the miserable survival and cure rates for this stage. Collective results of operation alone for N2 disease demonstrate 5-year survival rates ranging from 14% to 30%.^36,94 Using mediastinoscopy and positron emission tomography (PET), mediastinal lymph node involvement may be detected before thoracotomy, allowing the patient with N2 disease to be selectively treated with combination therapy in a neoadjuvant fashion, before resection. A number of investigators have looked at such neoadjuvant therapy for stage IIIA disease, and several small randomized trials of this approach have shown dramatic improvements in survival.^16,75,84,95 In many centers, this approach has become the standard care for patients with “limited” (usually defined as less than multistation and nonextranodal) N2 disease.

Although patients with IIIB (N3) disease would at first glance appear to be unlikely to benefit from combined-modality therapies including surgery, since N3 lymph nodes are out of the field of an ipsilateral mediastinal lymphadenectomy, several studies have suggested that neoadjuvant treatment followed by surgery for IIIB (N2) disease may also be of benefit.

Perhaps more theoretically appealing is the potential benefit of neoadjuvant treatment for stage IIIB (T4) NSCLC. Here the possibility of rendering a large, invasive tumor, which is unlikely to be resected with clean margins, “resectable” by prior treatment with chemotherapy, radiation, or both has obvious appeal. Nonrandomized studies have suggested that this approach may be effective, and it has been adopted by many surgeons in fit patients who can withstand an aggressive treatment plan.

To summarize, then, neoadjuvant treatment of NSCLC is most often currently considered in patients with disease that is stage IIIA or greater. The available evidence suggests that neoadjuvant treatment is more advantageous than postoperative adjuvant treatment, thus placing a premium on preresectional evaluation of mediastinal lymph nodes by mediastinoscopy. There is no strong consensus on the value of postoperative adjuvant treatment even in stage IIIA and beyond, but it is not unreasonable to consider this in generally healthy patients, preferably within a research protocol. The positive experience with neoadjuvant treatment in locally advanced cases has led some to suggest that similar benefits may be seen by applying neoadjuvant treatment to earlier-stage disease. Trials investigating this concept are ongoing. Postoperative chemotherapy for stage II NSCLC is an emerging indication, requiring close discussion between the surgeon, oncologist, and patient.

Interpretations of the results of clinical trials in the field of combined-modality therapy must be carried out with great care. Many of the trials evaluating the aforementioned treatment regimens involved patients with bulky N2 disease but also included other patients with locally advanced disease, such as disease involving the chest wall (T3) or locally invading the mediastinum. Including patients with T3N0 chest wall disease would tend to improve overall survival in a trial because these patients, as a rule, have a better long-term outlook than patients with N2 disease, and they now are staged as IIB. Other trials have included patients with stage IIIB disease (T4 primary tumor, contralateral
mediastinal lymph nodes). Importantly, not all trials included rigorous staging of the mediastinum in the form of a mediastinal lymph node dissection.

A variety of therapeutic modalities have been compiled in review articles. One review by Brundage and colleagues¹⁵ included 441 phase II studies and 108 phase III reports that enrolled stage III patients between 1966 and 1993. Review of the literature demonstrated significant diversity in research practices. Analysis of trials for stage III management found five major types of variation between studies: selection of control arms, selection of study investigational arms, choice of eligibility criteria, outcomes measures selected for study, and magnitude of benefit sought in the primary outcome measure. This diversity of research studies has made it increasingly difficult to definitively decide on the best treatment option.

As studies are reviewed, what parameters are truly reflective of clinical benefit to the patients should be considered. Median survival is not a particularly useful indicator. The aim as clinicians is to increase the number of long-term survivors. Also, another parameter to consider is response to induction therapy and adjuvant therapy. Grading of response, according to Milano and colleagues,⁴⁸ is a valid parameter to evaluate standard regimens and novel drug associations. Repeat mediastinoscopy after induction therapy also has been suggested for restaging of the extent of the disease after the initial therapy.^53,58

Neoadjuvant radiation therapy is at face value an appealing concept. Preoperative radiation has a good chance of reducing tumor size, facilitating resection, or possibly even downstaging the tumor. It may also minimize seeding of tumor cells by surgical manipulation and sterilize the tumor bed, but evidence for this is lacking. Tumor cells are particularly sensitive to radiation therapy at higher oxygen tensions; therefore preoperative use would be expected to be most effective on the highly vascularized peripheral border of the invasive tumor.³¹ Depending on the total dose of radiation delivered preoperatively, investigators—including Sherman⁸⁰ and Shields⁸¹ and their associates as well as Warram⁹⁸ and Bloedorn¹¹—have reported a range of 20% to 50% of patients having no persistent tumor or only microscopic disease in the resected specimen.⁶⁷ One of the problems with preoperative radiation therapy, however, is a tendency for significant fibrosis and loss of distinct tissue planes, making resection challenging. Further, the doses required to obtain a response when a single agent is given may be associated with decreased healing of the bronchial stump, resulting in bronchopleural fistulas.

The first use of neoadjuvant radiation therapy dates back to 1955, when Bromley and Szur¹⁴ at the Hammersmith Hospital used a dose of 45 Gy before surgical resection (Table 117-1). Sixty-six of 573 patients were resected. At operation, no viable tumor was found in 47% of the patients. Ten of the patients died of complications within the first month, and only two patients were alive 5 years later. Subsequently in 1964, Bloedorn¹¹ used a radiation dose of 60 Gy preoperatively to treat 109 patients with presumed unresectable lung cancer. The postoperative mortality was approximately 35%, and the 1-year survival approached 20%. Bloedorn¹¹ reported that tumor sterilization rates of 54% at the primary site and 92% at the mediastinal lymph nodes were obtained, but the status of the mediastinal nodes was not histologically documented before initiation of the preoperative regimen. Thus it was suggested that irradiation at 50 to 60 Gy is able in some cases to convert a surgical resection from incomplete to complete.

Other phase II trials included the following: a report from the Harvard Medical School evaluated 53 patients with marginally resectable NSCLC treated with 30 to 40 Gy of preoperative radiation therapy followed by resection and postoperative radiation therapy. Thirty-eight (72%) patients were resected. Sherman and colleagues⁸⁰ reported that the 5-year survival rate for the 38 resected patients was 27%, whereas it was 18% for all 53 patients. At Rush-Presbyterian–St. Luke’s Medical Center, Reddy⁶⁷ and associates reported on 74 patients with clinical stage III NSCLC who were treated with a 40-Gy preoperative radiation dose to the primary tumor in the lung and regional lymph node areas. Fifteen patients (20%) did not undergo operation because of tumor progression, patient refusal, or death. At the time of surgery, two patients had histologically negative specimens, nine had microscopic disease only, and 37 had gross residual disease. The 5-year and recurrence-free survival rates for the entire group were 20% and 24%, respectively. Patients with a complete pathologic response had a recurrence-free survival rate of 53% at 5 years, whereas only 17% of those with gross residual disease at surgery remained recurrence-free at 5 years. Half the patients with clinically uninvolved nodes were living recurrence-free at 5 years, compared with only 20% of those with N2 disease.

In the 1960s and 1970s, the Veterans Administration and the National Cancer Institute performed two large-scale trials that
randomized patients to immediate surgery versus preoperative 40- to 50-Gy radiation. In the Veterans Administration study reported by Shields and associates,⁸² 331 male patients with biopsy-proven bronchial carcinoma were randomized. No statistically significant increase in survival was noted in the pretreatment group (12.5% versus 21.0%). In fact, the survival rate in the preoperative treatment group was significantly lower during the first 12 postoperative months than in those patients who underwent resection only.⁸¹

In the National Cancer Institute study (1969), patients thought to be resectable at the time of diagnosis were randomly assigned to receive either immediate surgery (n = 278) or preoperative irradiation followed by surgery (n = 290). The 3-year survival rates for these two groups were nearly identical. At 5 years, Warram⁹⁸ reported that the survival rate was 14% after preoperative radiation therapy and 16% after immediate surgery. The preoperative radiation therapy was believed not to improve the resectability or survival rates in either study. Long-term survival was not improved even though local control was enhanced. Operative mortality was 12% in both groups.

These trials certainly did not identify a benefit of neoadjuvant radiation therapy. However, since they lacked pretreatment histologic staging and there was significant variation in the amount of radiation therapy delivered, as well as excessively long intervals between radiation therapy and surgery, the issue was not put completely to rest. Furthermore, some of the trials did not exclude patients with small-cell histology.

By the late 1980s, an interesting phase II clinical trial was started by the Lung Cancer Study Group (LCSG) to determine whether either preoperative radiation or chemotherapy was sufficiently active and safe to merit further work. Each arm was to be evaluated for its ability to induce an approximate 15% incidence of complete histologic clearance. Sixty-seven patients with stage III NSCLC were enrolled to receive either preoperative MVP chemotherapy (cisplatin, 120 mg/m²; mitomycin, 8 mg/m²; vinblastine, 4.5 mg/m²) or preoperative radiation therapy (44 Gy). All patients had systematic surgical staging of the mediastinum. Radiologic response to treatment was virtually identical for the two approaches (54% versus 48%), with 29 of the 57 available patients achieving objective responses. Twenty-three (40%) of the 57 patients eventually underwent complete tumor resection. Wagner and associates⁹⁷ recorded that median survival for the entire group was 12 months, with a 4-year survival rate of 27%. Thus, despite rigorous staging and excellent quality control, a more favorable stage III subset had a disappointing result.

Pastorino,⁵⁷ Cullen,²³ Vokes,⁹⁶ and Kris³⁹ have shown that neoadjuvant chemotherapy may also offer benefits in the management of locally advanced NSCLC (stage IIIA disease). It has the potential of reducing tumor size, may prevent tumor progression caused by perioperative immunosuppression and release of growth factors related to wound healing, and theoretically may eliminate clinically occult micrometastases. As suggested by a mathematical model, chemotherapy can eradicate a neoplastic subclone if the number of cells is <10⁶. Further, neoadjuvant chemotherapy gives the oncologist an opportunity to assess individual tumor sensitivity. The degree of response can be assessed histologically from the surgically resected specimen. From these results, Pujol and associates⁶⁴ suggest that it may be possible to individualize postsurgical treatment.

Many investigators—including Tonato,⁹² Shepherd,⁷⁹ Ginsberg,³⁶ and Ihde³⁸—even before the publication of randomized trials on the issue, had taken the position that neoadjuvant chemotherapy is beneficial in patients with otherwise inoperable stage IIIA NSCLC. The long-term survival in phase II studies has reached 18%, compared with 9% of historic control subjects, as reported by Pastorino.⁵⁷ Although many centers have essentially adopted neoadjuvant chemotherapy as their standard management of stage IIIA (N2) NSCLC on the basis of these data and the available small phase III studies, many remain skeptical about neoadjuvant chemotherapy because of what they consider the weaknesses of these trials. Many of these studies have poorly defined eligibility criteria, poor pretreatment staging systems, or lack of adequate numbers to draw a definitive conclusion. However, efforts continue to define the role of neoadjuvant chemotherapy in the management of stage IIIA disease (Table 117-2).

Perhaps the prototype phase II trial of neoadjuvant therapy was begun by investigators at Memorial Sloan–Kettering Cancer Center in 1986 and reported by Kris⁴¹ and Martini⁴⁶ and their associates. A preliminary phase I/II clinical trial conducted between 1984 and 1986 comprising 41 patients with bulky N2 disease set the stage for this trial. The preliminary trial demonstrated that a treatment program consisting of induction chemotherapy, operation, and postoperative radiation therapy (reserved for patients with residual disease at thoracotomy) yielded acceptable and perhaps improved median survival and long-term survival for some patients.² In the subsequent study, again trying to keep the patient population as homogeneous as possible, only patients with bulky N2 disease were entered. Patients received two to three cycles of mitomycin C, vindesine or vinblastine and high-dose cisplatin (120 mg/m²). Even if no radiologic response could be shown, they underwent surgery 4 to 6 weeks later as long as no progression of disease occurred. This experimental design, with a few variations, became the model for numerous other studies that followed, including those at the cooperative group level. A 77% “major radiographic response” rate to the chemotherapy occurred (105 of 136 patients), with a 10% (13 of 136) incidence of complete response. Of these patients, 98 had a thoracotomy. Complete resection was achieved in 65% of patients, and no histologic evidence of tumor was found in 19 (21.3%) completely resected patients. Median survival for the completely resected group was 27 months, with a 3- and 5-year survival rate of 44% and 26%, respectively. For the entire group, overall survival at 1, 3, and 5 years was 72%, 28%, and 17%, respectively. In patients in whom a pathologic complete response was achieved, survival was 95%, 71%, and 61% at 1, 3, and 5 years, respectively. Median survival was 64 months in this group. Also of note, 78 of 136 patients received postoperative radiation therapy to the mediastinum.

This trial demonstrated that with a combined chemotherapy and surgery approach, survival in stage IIIA patients with the worst prognosis (those with bulky N2 disease) could approximate that observed in patients with the best prognosis (microscopic N2 disease discovered at the time of thoracotomy in a patient in whom it was not suspected).^40,46,47 This represents a
significant advance when one realizes that with surgery alone, there is essentially a 0% survival at 3 years in the group of patients with bulky mediastinal nodal disease and less than a 10% rate of resectability.