Presurgical Cytoreduction of Oral Cancer Using Intra-arterial Cisplatin and Limited Concomitant Radiation Therapy (Neo-RADPLAT) FREE

Traditionally, advanced oral cancers have been treated by surgery followed by radiation therapy. A major drawback of this primary surgical approach is that extensive loss of tissue is often required to adequately resect the mucosal disease. This may involve the tongue, mandible, suprahyoid musculature, soft palate, and pharyngeal musculature. Consequently, important bodily functions such as deglutition, mastication, and speech are often compromised, sometimes severely. Furthermore, the rates of local disease control and survival continue to be less than optimal and little improvement has been achieved despite refinements in surgical techniques and radiation therapy.

To minimize the functional disadvantage of important oral tissue loss necessitated by extensive surgical resections, we conducted a pilot study using a 2-component protocol: first, an induction therapy consisting of intra-arterial (IA) infusion of cisplatin and concomitant radiation therapy, for which the total radiation dose was less than the standard amount when used as a single modality; then, surgery limited to a tumor nidusectomy (a limited resection of the residual tumor bed). Nicknamed neo-RADPLAT—with neo denoting a modification of the original RADPLAT (cisplatin plus radiation) protocol—the overall strategy was to use the first component to achieve cytoreduction of disease in the primary site and thus minimize the tissue resection required in the second component of therapy. The preliminary experience of this trial was reported in abstract form.¹ We now report our findings on the complete cohort of 25 patients, all of whom have had a minimum of 2 years of follow-up.

The analysis included 25 patients treated at the University of Tennessee Health Science Center, Memphis, between October 1995 and June 2000. All patients had stage II, III, or IV squamous cell carcinoma (17 patients had cancer of the oral cavity and 8 had cancer of the oropharynx). Prior to therapy, all patients were presented at the weekly multidisciplinary treatment planning conference. Based on conventional treatment approaches, each patient was considered to be a candidate for primary surgery and likely to require postoperative radiation therapy. No patient was considered to be a good candidate for primary radiation therapy owing to the size and/or site of the primary lesion. The most important deciding factor for patients to undergo the protocol rather than primary surgery was the volume and extent of tissue that would be removed at the time of primary surgery. Thus, for each patient, there was substantial risk of a need for extensive resection with loss of organ function if primary surgery was used. All patients were required to sign an informed consent statement approved by the institutional review boards of the institutions involved.

The median age of the patients, 18 men and 7 women, was 60 years (range, 18-71 years). Table 1 outlines the distribution of patients by T and N classification. Two patients had stage II, 10 had stage III, and 13 had stage IV disease; there were 5 patients with T2 and 20 with T3 disease; and 12 patients had N0, 9 had N1, and 4 had N2 neck disease (2 patients had N2a, 1 had N2b, and 1 had N2c disease). The patient with N2c disease also had N1 disease on each side of the neck. Among the 17 patients whose primary site of disease was the oral cavity, 11 had oral tongue cancer, 5 had retromolar trigone cancer, and 1 had primary cancer arising in the floor of mouth. Among the 8 patients with disease arising in the oropharynx, 3 had a base of tongue, 4 a tonsil, and 1 a soft palate primary cancer.

The neo-RADPLAT protocol consisted of chemotherapy using cisplatin, which was rapidly infused intra-arterially through a transfemoral catheter strategically placed in the external carotid artery to selectively infuse the primary tumor. Details of the technical aspects of the catheterization and infusion have previously been described.² Based on an earlier phase 1 trial in which the maximum tolerated dose of cisplatin was determined,³ the quantity of cisplatin infused was 150 mg/m² on days 1, 8, 15, and 22 of radiation therapy. Simultaneously with the IA administration of cisplatin, the neutralizing agent sodium thiosulfate was given intravenously at the rate of 9 g/m² over 30 minutes, followed by 12 g/m² over 6 hours. Thus, the protocol allowed the planned dosage of cisplatin to be 150 mg/m² per week, compared with 30 to 40 mg/m² per week in the more conventional intravenous chemotherapy protocols.

Radiation therapy consisted of fractionations of 2 Gy delivered once daily by external beam using standard portals of parallel opposed fields. Patients received 10 Gy/wk to a total dose of 50 Gy over 5 weeks. Both sides of the neck were included in the field as well as the supraclavicular region. With the exception of the posterior neck overlying the spinal cord, all areas received a total dose of 50 Gy over 5 weeks.

At 8 weeks following radiation treatment, patients underwent surgery on the basis of the results of computed tomographic scans, examination under anesthesia, and a tumor nidusectomy. The tissue excised from the primary site was then transected serially for frozen-section analysis. If there was evidence of residual tumor, a wide excision of the tumor bed was performed until clear margins were obtained. However, no further surgery was done if frozen section analysis revealed no residual disease. Neck dissection was also performed for patients who had evidence of residual nodal disease. The type of dissection was selective, with removal of neck levels I, II, and III for oral cavity primary tumors and of neck levels II, III, and IV for oropharyngeal tumors.

Follow-up analysis was done for persistent disease, local disease recurrence, regional recurrence, distant metastases, and survival. Estimates of locoregional treatment failure and of overall and cause-specific survival at 5 years were plotted using the Kaplan-Meier method.

Twenty-one patients (84%) received the 4 chemotherapy infusions prescribed, 1 patient received 3 infusions, 1 received 2 infusions, and 1 received only 1 infusion. With regard to radiation therapy, all patients except 1 received a total dose of 50 ± 2 Gy. The remaining patient received 70 Gy. Immediately following the first chemotherapy infusion, this patient experienced grade 3 neurotoxic effects (National Cancer Institute common toxicity scale). The decision was made to discontinue chemotherapy and increase the total dose of radiation therapy to 70 Gy over 7 weeks.

There were 26 toxic events of grades 3, 4, or 5 involving 19 patients (Table 2), among whom 16 had grade 3 mucositis. Of the 2 patients (8%) who experienced severe adverse events, one had grade 4 neutropenia and the other grade 5 cardiac toxicity. This patient died from cardiac myopathy 1 month following completion of therapy. Finally, the patient who experienced neurotoxicity had a transient polyneuropathy of cranial nerves VII and V.

Twenty patients (80%) had a complete response to chemoradiation in the primary site administered on the basis of pathological criteria (14 patients) or clinical assessment (6 patients) (Table 3). Following the completion of the chemoradiation therapy, 14 patients underwent examination under anesthesia and a nidusectomy. Among them were 9 patients who also had a selective neck dissection (1 bilateral) because of clinical evidence of residual nodal disease. Five patients had frozen section evidence of residual tumor in the primary site and thus had a conventional oncological resection following tumor nidusectomy.

Ten patients (79%) had a complete response to chemoradiation in the neck based on pathological criteria (5 patients) or clinical assessment (5 patients). One of these patients had a bilateral neck dissection, thus accounting for 11 necks at risk. Among the 11 neck dissection specimens, 3 revealed pathological nodes, all located in level II. There were 6 patients who did not undergo a nidusectomy of the tumor bed, and all of them had a clinically complete response. Included in this subset were 3 patients whose medical status had deteriorated and who became marginal surgical candidates. The remaining 3 patients refused the recommended procedure (Table 4, Table 5, and Table 6).

With a median follow-up of 56 months (range, 28-84 months), the 5-year estimates for overall survival, disease-specific survival, and locoregional control are 54%, 64%, and 74%, respectively (Figure 1 and Figure 2). Fourteen patients remain without disease, 6 have died of the disease, and 5 have died of other causes. The initial site was local in all 5 patients who had treatment failure. Among them, only 1 was successfully treated with salvage surgery but he subsequently died of a second primary cancer (of the piriform sinus). One had a simultaneous regional failure, and another had a simultaneous distant failure. Two of the 5 patients who died of another cause developed a second primary tumor.

The data from this study indicate that it is feasible and efficacious to treat patients who have intermediate (T2) or advanced (T3-T4) resectable carcinoma of the oral cavity, or advanced (T3-T4) resectable oropharyngeal carcinoma, with triple-modality therapy. Furthermore, based on the small number of patients in whom it was necessary to perform a wide (oncological) resection following chemoradiation therapy, there appears to be the added advantage of tissue preservation. Using preoperative IA cisplatin chemotherapy in an extremely high-dosage schedule (150 mg/m² per week) and concomitant radiation therapy in a total dose less than the conventional therapeutic amount, cytoreduction was achieved and excessive loss of tissue important for oral function was avoided.

The effectiveness of induction chemotherapy for head and neck cancer has been the subject of investigation for several years. Randomized trials have not shown any increase in survival with this approach, although the subset of patients who had a complete response has benefited because of improved survival and organ preservation.^4- 5 Unfortunately, it has been possible to achieve a complete response in only 30% to 40% of patients with this approach. Another problem with induction chemotherapy when used sequentially with radiation therapy has been the necessity to use in the latter total doses that are therapeutic, typically in excess of 70 Gy over 7 weeks.^5- 6 To a lesser extent, induction chemotherapy has also been tested with concomitant radiation therapy, but always with a therapeutic total dose radiation or with surgery.^7- 9 In this setting, there are significant concerns related to toxic effects on the tissue. Furthermore, the 2 combined modalities in which maximum doses are given are sufficiently potent to serve as the definitive therapy even for advanced disease. Thus, our protocol's strategy of concomitant chemoradiotherapy is unique with regard to the dosages used. We purposefully limited the total dose of radiation to minimize the risk of long-term radiation effects on oral soft tissue and bone and to allow safer surgical intervention. We also believed that it was possible to spare salivary tissue with this strategy. In this pilot study, we chose to use 50 Gy in anticipation that some salivary function would recover. Furthermore, although not routinely prescribed for the patients in this series, there existed a theoretical advantage for the use of amifostine and its potential for stimulating residual salivary gland tissue in this setting.

Treatment delivery and toxicity are significant concerns when the protocol involves concomitant chemoradiotherapy. Defining 50 Gy as a complete total dose for radiation therapy and 3 or more IA infusions of 150 mg/m² of cisplatin as the minimum dose for chemotherapy, all of the patients received the minimum prescribed total dose of radiation and only 2 patients had fewer than 3 intra-arterial infusions. This 92% rate of patients who received the complete induction chemoradiation regimen is in keeping with our larger experience using IA chemoradiation as the definitive therapy for advanced disease for which the total dose of radiation therapy is 70 Gy.¹⁰ Grade 3, 4, and 5 toxicity was encountered 26 times in 19 patients. However, most of the toxic events were related to grade 3 mucositis, defined as confluent desquamation. This toxic effect is common among patients receiving radiation therapy alone or with concomitant chemotherapy. Excluding mucositis, the toxic events were mainly hematologic and gastrointestinal, and were all successfully treated. The grade 5 toxic event, death, was due to a cardiomyopathy that developed 1 month after completion of therapy and may not have been directly related to the treatment. The single case of neurotoxicity involved a patient who developed transient cranial neuropathy of nerves V and VII immediately after the first IA infusion. We believe that this was a direct effect of the cisplatin infused locally. The patient recovered spontaneously and tolerated a full total dose of radiation therapy without any more chemotherapy. Although cranial neuropathies related to IA chemotherapeutic infusions have been reported, this was only the second patient in our experience with more than 400 patients undergoing IA cisplatin chemotherapy. Our data indicate that life-threatening toxic events can occur with the induction chemoradiation approach used in the study. However, such events are uncommon and well within the accepted standards of the risk-benefit ratio faced by patients with advanced head and neck cancer undergoing combined modality treatment.

The complete response rate to the induction chemoradiation schedule used in the study was 80% in the primary site and 79% in the regional nodes. These results are not significantly different from the complete response rates observed in the series of patients receiving definitive therapy with RADPLAT.¹⁰ Despite the use of a decreased total dose of radiation therapy, the observed high rate of complete response suggests that maximum doses of radiation may not be necessary when boosted by chemotherapy. This observation raises the question of whether maximum therapeutic doses are needed when used in the context of combined modality therapy, particularly if it involves 3 components. It is possible that multimodality therapy will allow to lower doses for each modality, and thereby minimize short and long-term toxic effects on tissue. In our protocol, we reduced the dose of radiation and the extent of surgery. Using less chemotherapy may also be explored in this setting.

Although it is well known that therapeutic doses of radiation must exceed 65 Gy within 6 to 7 weeks to be effective as a single-modality treatment, the rationale for its use in this study was based on the synergistic action of cisplatin given concomitantly. Preclinical and clinical studies have demonstrated this synergistic effect.^11- 12 Furthermore, the extremely high dosage of cisplatin (150 mg/m² per week) delivered intra-arterially and directly into the tumor bed added a direct cytotoxic boost. We believe that the added effect of the cisplatin made it possible to limit the total dose of radiation without compromising treatment efficacy.

This novel protocol also attempts to capitalize on the synergy provided by combined modality therapy by incorporating a third modality, surgery. If it had been used alone or in combination with one other modality, it is likely that the surgical approach used in our study would have been suboptimal for most patients. However, surgery as a component of a triple-modality strategy appears to be feasible and advantageous. Thus, when limited surgery is used sequentially as the third component following concomitant chemoradiation, the data support this concept, particularly for patients who had no evidence of residual tumor in the nidusectomy specimen.

A key issue related to this protocol is whether oncological resection is feasible in patients who did not have a complete response histopathologically demonstrated following tumor nidusectomy. Whereas the local control rate was 92% in the subset sensitive to chemoradiation, 5 patients resistant to this cytoreduction therapy required the added step of a wide resection. This resulted in a local control rate of 60% among this subset.

A weakness of the study was the lack of objective measurements regarding long-term toxic effects and quality of life. Although salivary measurements were performed for some of the patients in this pilot study, there were insufficient data to draw any firm conclusions. However, although there were complaints from a few patients related to chronic xerostomia, such complaints were much fewer than expected. It would be important to incorporate such end points in future studies. With regard to wound healing, there were no serious complications or fistulas, although at least 3 patients had minor problems with wound breakdown and healing by the process of secondary intention.

Corresponding author and reprints: K. Thomas Robbins, MD, Division of Otolaryngology, Southern Illinois University, PO Box 19662, Springfield, IL 62794-9662 (e-mail: ent@siumed.edu).

Submitted for publication April 30, 2003; final revision received July 28, 2003; accepted August 19, 2003.

This work was supported by a grant from Baptist Memorial Hospital Foundation, Memphis, Tenn (Dr Robbins).

This work was presented in abstract form at the annual meeting of the American Head and Neck Society; May 2, 2003; Nashville, Tenn.