Forty-six patients (43 men and 3 women), aged from 39 to 75 years (mean 58,3, median 42,1), with previously untreated squamous cell carcinoma of the upper aerodigestive tract were treated between November 1993 and December 2000. The pre-treatment characteristics of the patients and the site and stage of the tumour, according to the 4^th edition of the TNM used at that time 12, are reported in table 1 and table 2. The T2 patients were submitted to this protocol owing to general conditions that contraindicated surgery or because they refused surgery. Patients were required to sign informed consent approved by the Ethical Committee of the Institution.

Patients underwent a complete clinical and laboratory examination including measurement of haematologic (complete blood cell and platelet count, prothrombin time, partial thromboplastin time), hepatic (blood bilirubin, transaminases and alkaline phosphatase level)and renal (blood urea nitrogen, creatinine and electrolytes level, urinanalysis, 24 hours creatinine clearance) parameters, as well as general metabolic functions (blood glucose and lipids level), chest radiography and cervical vessel ultrasonography. The disease was staged by physical examination, panendoscopic examination, biopsy, CT scan and/or MRI.

All patients received intra-arterial (i.a.) infusions of carboplatin, dissolved in 80 mL of 0,9% saline and infused over 10–15 min by a battery-operated pump (Medrad Mark 5 Plus) with an infusion velocity of 6–8 mL/min. An angiographic catheter (Glidecath Radicofocus 5F Terumo) was introduced percutaneously under local anesthesia (10 mL of 2% lidocaine solution) into the common femoral artery, according to the Seldinger technique. The catheter tip was placed in the external carotid artery, under radiographic control. Diagnostic transfemoral carotid arteriography was performed and the most suitable branch of the artery, providing the main blood supply to the tumour, was selectively catheterised. The catheter was removed immediately after the infusion. A simultaneous intravenous (i.v.) infusion of 250 mL of 0,9% saline was performed. Dexamethasone (4 mg) and ondansetron (1 g) dissolved in 100 mL of 0,9% saline were administered i.v. 1 hour before the infusion. All patients received three cycles of intra-arterial chemotherapy, administered every two weeks. The planned dose of carboplatin per cycle was 300 mg/m² for the first cohort (3 cases) and 325 mg/m² for the second cohort (3 cases); subsequently, as no significant toxic effects were observed, the remaining patients (40 cases) received a dose of 350 mg/m² of carboplatin per cycle. Two weeks after completing chemotherapy, patients were restaged and subsequent treatment was decided upon on the basis of their objective response. Complete responders or partial responders had radiotherapy; non-responders with resectable disease underwent surgery on the tumor site and/or on the neck nodal metastases followed by radiotherapy; non-responders with unresectable disease underwent palliative radiotherapy. Radiation was delivered using conventional fractionation (1,8 – 2 Gy per fraction, five days per week). The total dose was 66 – 70 Gy for PTVI (planning target volume including tumour and encompassing all clinically positive nodes) and 50–56 Gy for PTVII (planning target volume including all anatomical regions at risk of subclinical malignant disease).

Toxicity was evaluated according to WHO criteria 13 24 hrs after each drug administration. This included: assessment of haematologic, hepatic and renal parameters; assessment of gastrointestinal, otoneurological, cardiac and respiratory function, and the search for signs and symptoms of local toxicity. Response to therapy was assessed two weeks after the last administration of the drug by physical examination, endoscopy, imaging (CT scan and/or MRI) and biopsy, when necessary. A complete response (CR) was defined as the complete disappearance of all demonstrable lesions. A partial response (PR) was defined as a decrease of 50% or greater in the sum of the products of the largest perpendicular diameters of all measurable lesions. No response (NR) was defined as a regression of < 50% in total tumour size or as stable or progressive disease.

The primary efficacy end point was overall survival (i.e, death as a result of any cause). Survival was measured from the start of treatment date to the date of death or the date when the patient was last known to be alive. Yearly estimates were calculated using the Kaplan-Meier method. Secondary efficacy end points were CR rates, loco-regional tumour control and disease-free survival. Time to loco-regional failure was measured from the start of treatment date and the date of disease relapse, the date of disease-related death, or the date when the patient was last known to be alive and disease-free. Loco-regional failure, distant metastases, or disease-related death were all considered failures for disease-free survival. Yearly disease-free survival rates were estimated with the Kaplan-Meier method.

Factors influencing overall and disease-free survival were analyzed and compared by the log-rank test. P values less than 0.05 were considered significant.

One hundred and thirty eight super-selective transfemoral intra-arterial infusions of carboplatin were performed with no significant complications.

Carboplatin treatment was well-tolerated by the majority of the patients. No instances of gastrointestinal (nausea, vomiting, diarrhoea), renal, cardiac, respiratory and otoneurological toxicity were reported. Cases of systemic and local toxicity are listed in table 3.

After completing chemotherapy, 16 patients had a CR (35%), 20 patients had a PR (43%) and 10 patients had NR (22%). The 36 patients with a complete or partial response underwent radiotherapy, after which 26 patients achieved a complete response. Among the other 10 patients with persistent or recurrent disease in the neck, 5 underwent neck dissection and achieved a complete response and 5 refused any further treatment. Seven of the 10 NR patients underwent neck and primary site surgery, followed by radiotherapy. After this treatment, all 7 patients achieved a complete response. Of the remaining 3 non-responding patients, one refused any further treatment and two had palliative radiotherapy. Upon completing all treatments, 38/46 patients (83%) achieved a complete response: the 26 patients (CR) treated with induction chemotherapy and radiotherapy; the 5 (PR) treated with induction chemotherapy, radiotherapy and neck surgery and the 7 (NR) treated with induction chemotherapy, primary site and neck surgery and radiotherapy. (Fig. 1)

No patients were lost to follow-up. The median follow-up time, measured from the start of induction chemotherapy, was 51.3 months (range: 5 to 136 months).

The current status of the patients is as follows: after completing the treatment protocol, the 8 patients with persistent or recurrent disease (3 from the NR group and 5 from the PR group) have died of the disease; 13 of the 38 CR patients are alive and disease-free, 3 have died of distant metastases (2 pulmonary and 1 hepatic) and 3 due to a second primary tumour (2 lung cancers and 1 renal cancer), which occurred after a mean period of 29 months. The tumour recurred in 19 patients after a mean period of 25 months; 9 on the neck and 10 at the primary site. Five of the 9 patients with tumour recurrence on the neck underwent neck dissection; of these 2 are still alive and disease-free and 3 have died due to the persistence of the tumour; the other 4 received systemic chemotherapy. Of these 1 is alive and free of disease while 3 have died of the disease. Of the 10 patients in whom tumours recurred at the primary site, 5 underwent palliative systemic chemotherapy and died of the disease, 2 received only palliative medical treatment and died of the disease and 3 underwent salvage surgery. Of these, 2 are alive and disease-free and 1 has died of pulmonary metastases. In conclusion, after a 5-year follow-up, 18 patients (39%) are alive and disease-free, 3 (6,5%) have died due to a second primary tumour and 25 (54,5%) have died of the disease (Fig. 2).

The percentage of organ preservation was 83%, as 8 of the 46 patients underwent salvage surgery at the tumour site.

The probability of five-year overall survival was 50% (Fig. 3). The main factor influencing overall survival was the response to induction chemotherapy: CR patients showed statistically significant better survival rates compared to PR or NR patients, even if 12 (40%) of these patients were successfully submitted to salvage surgery (Fig. 4).

Five-year disease-free survival, calculated on the 38 patients who were free of disease after completing all treatments, was 49,8% (Fig. 5). In this case, the main factor influencing disease recurrence or the appearance of distant metastases was patients' neck status: patients with no metastatic lymph-nodes or patients whose metastatic lymph-nodes regressed completely after induction chemotherapy had a better disease-free survival rate than patients with positive nodes who had a partial or no response to chemotherapy (Fig. 6).