- Email: [email protected]
Composite resection with mandibulectomy in the treatment of posterolateral oral cavity and lateral oropharynx squamous cell carcinoma
Composite Resection With Mandibulectomy in the Treatment of Posterolateral Oral Cavity and Lateral Oropharynx Squamous Ceil Carcinoma Jean-Louis Lefebvre, MD, Bernard Vankemmel, MD, Bernard Prevost, MD, Etienne Buisset, MD, Bernard Coche-Dequeant, MD, Jean Ton Van, MD, Thierry Oszustowicz, MD, LilleC6dex,France
From 1972 to 1987, 403 patients underwent a composite resection consisting of segmental mandibulectomy and neck dissection for the treatment of bueco-pharynx squamous cell carcinoma (303 patients had postoperative radiotherapy [XRT], 29 patients had preoperative XRT, and 100 patients had tissue salvage performed after XRT). Of the 303 patients who received preoperative XRT, 32 were clinically staged T2, 149 were staged T3, and 122 were staged T4; 194 of the 303 patients were staged NO. In the 100 patients who experienced relapse and who required "salvage surgery," the restaging found 20 patients staged T2, 39 staged T3, and 41 staged T4; 73 of the 100 patients were staged NO. The patients' mean postoperative stay was 15 days for those who had prior surgery and 21 days for those who had salvage surgery. With a minimum follow-up of 5 years, locoregional recurrences and postoperative death occurred in 86 of 2 7 4 patients (31%) in whom surgery and postoperative XRT were performed; in 17 of 29 patients (59%) who had preoperative XRT performed; and in 61 of 100 patients (61%) who had salvage surgery performed. In addition, in terms of functional results, only 61% of patients ( 2 0 6 ) were able to maintain normal speech function, and only 23% ( 9 1 ) were able to maintain normal oral feeding. Of the 403 patients, the 5-year survival rates were 33% for the overall population, 42% for patients with postoperative XRT, 16% for patients in whom operation was performed after preoperative XRT, and 17% for patients who underwent salvage surgery.
ince its description by Hayes Martin in the 1940s, resection with segmental mandibulectomy Sand composite en-bloc radical neck dissection has been the procedure of choice in the surgical management of cancer of the posterolateral part of the oral cavity (PLOC) or of the lateral oropharynx (LO). In some cases, either due to mandible invasion or to the need for a proper surgical access for tumor resection, ablation of the posterior part of the mandible (including the ramus) is required. Such surgery has undoubtedly improved local control, but with the subsequent addition of some morphologie and/or functional sequelae. At Centre Oscar Lambret (Northern France Comprehensive Cancer Center), we often encounter buccopharyngeal tumors; approximately 200 patients with squamous cell carcinoma of the oral cavity, the oral tongue, or the oropharynx are referred to our institution yearly. In about half of these patients, the tumor is located in the PLOC or LO. According to our therapeutic protocols, tumors of the retromolar trigone, tumors of the very posterior part of the floor of the mouth (reaching or invading the mandible and for which a marginal mandibular resection would not be sufficient), tumors of the junctional area of the tongue (junction between the anterior pillar and mobile tongue), tumors of the tonsillar fossa (invading pterygoid muscles or the ramus or the glossotonsillar sulcus), and tumors of the glossotonsillar sulcus, itself, are treated by composite resection and postoperative radiotherapy. Other tumors of the PLOC or LO are treated by definitive radiotherapy (with neo-adjuvant chemotherapy for advanced stages), with surgery in reserve for salvage, if necessary. Our experience over a 15-year period is reviewed for this presentation.
PATIENTS AND METHODS Population: From 1972 through 1987, 403 patients were successively treated by surgery consisting of composite resection of the buccopharynx with segmental mandibulectomy and neck dissection (the so-called "commando" operation). There were 376 men with a mean age of 53 years (range: 31 to 82 years) and 27 women with a mean age of From the Departmentsof Head and Neck Surgery (J-LL, BV, EB, 59 years (range: 35 to 83 years). In 190 patients, the JTV), Radiotherapy(BP, BC-D), and Anesthesiology(TO), Centre primary site was the PLOC, that is, the retromolar triOscar Lambret,LilleC&lex,France. gone, the junctional area of the tongue, and the posterior Requests for reprints shouldbe addressedto Jean-LouisLefebre, part of the floor of the mouth. In 213 patients, the priMD, CentreOscar Lambret,Northern FranceComprehensiveCancer mary site was the LO, that is, the tonsillar fossa and Center, B.P. 307, 59020 LilleC&lex,France. Presentedat the 39th AnnualMeetingof the Societyof Head and pillars, and the glossotonsillar sulcus. Neck Surgeons,Los Angeles,California,March 18-21, 1993. Three hundred three patients (75%) were previously THE AMERICAN JOURNAL OF SURGERY VOLUME166 OCTOBER1993 435
LEFEBVRE ET AL
TABLE I
TABLE II
Results of Head and Neck Cancer Staging Criteria* Applied to Patients Who Received Preoperative Radiotherapy Stage of Nodal Involvement
T1
T2
T3
NO N1 N2a N2b N2c N3
----. --
23 5 4 --
101 26 18 3
Total no, of Patients
--
.
Total No: of Patients
Stage of Nodal Involvement
T1
T2
T3
T4
194 57 33 10
NO N1 N2a N2b N2c N3
I ---. --
i7 3 ---
29 5 4 -1
27 4 6 1 . 3
73 12 10 1
1
70 26 11 7 . 8
149
122
303
Total no. of Patients
--
39
41
100
. --
32
Results of Head and Neck Cancer Staging Criteria* Applied to Patients Who Experienced Relapse and Who Were Re.Staged for Salvage Surgery
T4
.
9
*American Joint Committee on Cancer (AJCC) criteria (1988).
.
. --
20
.
Total No. of Patients
4
*American Joint Committee on Cancer (AJCC) Criteria (1988).
/ f untreated. Surgery was the first therapeutic procedure, followed by postoperative radiotherapy (XRT) in 274 patients, whereas, in 29 patients, surgery was performed after preoperative XRT. In contrast, 100 patients (25%) had had previous treatment consisting of full-dose XRT and underwent operation(s) for local (with or without regional) recurrence. Clinical staging was evaluated according to the criteria recommended by the American Joint Committee for Cancer (AJCC, 1988). For accurate diagnosis, the staging was made at the time of decision-making for previously untreated patients. For patients who underwent operation(s) for a local relapse, this re-evaluation was made at the time of surgery. Using this classification, we found that for the 303 previously untreated patients (Table I), 271 patients (89%) had T3 and T4 lesions, and 109 (36%) had neck nodal involvement staged N1 to N3; 23 cancers were classified as stage II, 132 as stage III, and 148 as stage IV. For the restaging of the patients who experienced recurrence (Table II), we found that 80 patients (85%) had T3 and T4 lesions, and 26 patients (26%) had neck 436
THE AMERICAN JOURNAL OF SURGERY
Figure
1. C l a s s i c a l lip-and-chin-splitting i n c i s i o n used a s p a r t o f c o m p o s ite r e s e c t i o n w i t h m a n d i b u l e c t o m y .
nodal involvement staged N1 to N3; 17 cancers were classified as stage II, 37 as stage III, and 46 as stage IV. Surgical proeedutres: Before choosing to go ahead with the operation, all of the patients received a complete evaluation, including the following: panendoscopy, chest radiography, orthopantography, and classical preoperative biologic assessment; all had histologic confirmation of their disease, or of its recurrence (all were confirmed as cases of squamous cell carcinoma). When necessary, patients underwent nutritional support prior to surgery. Dental care was performed, as well. The evaluation of the tumor volume was assessed under general anesthesia (during panendoscopy), most often through a joint examination by a head and neck surgeon and a radiation specialist. Surgery was performed under general anesthesia with tracheotomy. We used a classical lip-and-chin-splitting incision (Figure 1), which recently has been replaced by a lip-splitting~-chin-sparing incision. The neck dissection was, in all cases, the first surgical procedure. In the case of neck nodal involvement staged NO or N1, we performed a modified dissection. In the case of palpable
VOLUME 166
OCTOBER 1993
COMPOSITE RESECTION IN BUCCO-PHARYNX CANCER
lymph nodal tumors over 3 cm in diameter (N2), a traditional radical neck dissection was performed; the spinal accessory nerve was preserved, whenever possible. Thereafter, the tumor excision was initiated: the mandible was sawed at the level of the mental foramen and removed after exarticulation of the temporomandibular joint, in continuity with the tumor. Surgical margins were assessed by frozen sections. During this period of time (1972 to 1987), we did not perform any mandibular repair operations. In contrast, we used, in 43 patients (11%), a pectoralis major myocutaneous flap for soft tissue repair. All previously untreated patients underwent a neck dissection, whereas 63 of the patients who experienced a recurrence underwent some surgery in the neck. The 37 patients who did not undergo neck dissection either had had prior surgery in the neck or had residual sclerosis without palpable lymph nodal tumors after XRT; these factors disinclined us to perform a neck dissection in these patients. Radiotherapy protocol: Patients who were treated for a local recurrence of cancer had received definitive XRT that delivered 60 to 65 Gy (10 Gy per week, 1 dose of 2 Gy per day) up to 70 to 75 Gy on residual disease or palpable lymph node involvement, using two opposite portals. Sources were a cobalt 60 machine and an electron beam for boosting. At 40 Gy, the field was reduced in order to spare the spine. Patients underwent operation at the time of recurrence or, in the case of persistent disease, 3 months after the completion of XRT. Patients who had preoperative XRT received 45 Gy over a 3-week period (15 Gy per week, 1 dose of 3 Gy per day) with similar fields, by means of a cobalt 60 machine. These patients underwent operation after a 2-week rest. Fifteen days after wound healing, patients who had postoperative XRT received 55 Gy on both the neck and pharynx--up to 70 Gy on positive margins or extracapsular spread areas; cobalt 60 and electron beam were used, which is similar to the protocol followed for definitive XRT. Finally, patients who underwent salvage surgery with myocutaneous flap received depilation XRT (if necessary) of 30 Gy electron beam (10 Gy per day) in the pectoralis area, over the 3 days preceding surgery. Statistical considerations: Survival rates were calculated according to the Kaplan-Meier method, and survival rates were compared by the log-rank test. Survival was evaluated from the day of pathologic diagnosis for previously untreated patients, and from the day of surgery for those patients who underwent salvage surgery. Comparison of effectiveness was statisticaUy analyzed with the x 2 test. RESULTS
Postoperative course: Eight patients (2%) died during the postoperative course: two of severe dyspnea (patients who had no tracheotomy but whose dyspnea very quickly became systematic), three of myocardiac infarction, two of multi-system organ failure, and one of massive aspiration. Thirty-one patients (8%) had a delay in wound healing. The mean tracheotomy duration was 12
TABLE
III
Actuarial Survival Rates by Percent for Patients at 3-Year, 5-Year, and 10-Year Follow-Up
Patient Groups
Percent of Percent of Percent of Total No. Patients Patients Patients of Patients at 3 Years at 5 Years at 10 Years
Overall population
403
48
33
17
Surgery and postoperative XRT
274
55
42
21
Preoperative XRT and surgery
29
24
16
16
Salvage surgery
100
23
17
NE
XRT = radiotherapy
days (range: 2 to 55 days), and two patients had delayed removal of the tracheotomy cannula (more than 3 weeks). The mean length of time the nasogastric feeding tube was in place was 13 days (range: 4 to 90 days); 22 (5%) patients had delayed removal of the feeding tube (more than 3 weeks), of whom 3 required a gastrostomy. For the overall population, the postoperative in-hospital stay was 17 days (range: 8 to 90 days): 15 days for patients who underwent prior surgery and 21 days for patients who underwent salvage surgery. Pathology: Frozen section revealed satisfactory results at first resection in 70% of patients; 25% needed further resection; and 4% were not evaluable, due to fragmentation of the specimen. On paraffin section analysis, surgical margins were negative in 76% of patients, suspicious in 17%, positive in 6%, and not evaluable in 1% (fragmented specimen). In neck dissection, there were 168 patients (46%) without nodal involvement, 80 patients (23%) with nodal involvement but without extracapsular spread, and 105 patients (29%) with extracapsular spread. In 111 patients (30%), there were 1 or 2 involved lymph node(s), and, in 80 patients (22%), 3 or more nodes were involved. Finally, in seven patients (2%) who underwent operation after definitive XRT, it was impossible to assess whether the lymph nodes were still evolutive or not. Long-term results: To date, with a minimum followup of 5 years, 133 patients (33%) have not demonstrated any cancer evolution nor second primary-site appearance. One hundred sixty patients (40%) demonstrated a cancer evolution but no second primary site, 86 patients (21%) demonstrated 1 second primary site, and 16 patients (4%) demonstrated both a cancer evolution and the presence of a second primary site. Local relapse occurred in 122 patients (30%) and nodal recurrence in 48 patients (12%). Both events occurred very quickly, with a mean time to occurrence of 9 months (range: 6 to 60 months, 98% within 24 months). Distant metastasis occurred in 38 patients (9%). Seventy-one (18%) patients developed a second primary site in the upper aerodigestive tract, 31
THE AMERICAN JOURNAL OF SURGERY
VOLUME 166
OCTOBER 1993
437
LEFEBVREET AL
TABLE IV " D i r e c t F a i l u r e s " Related to Therapeutic Protocol at a
Minimum Follow-Up of 5 Years
No. of Cases of Postoperative Death
Cases of Local and/or Regional Recurrence (*)
Totals (%)
Surgery and postoperative XRT
6
80 (66)
31
Preoperative XRT and surgery
--
17 (16)
59
2
59 (46)
61
Therapeutic Protocol Followed
Salvage surgery
*Number of patients who died due to uncontrolled local and/or regional recurrenceof cancer. XRT = radiotherapy,
TABLE V
Protocol Followed and Incidence of Negative Sequelae--Functional Results Protocol Followed and Patient Groups
No. of Patients (%)
Surgery and XRT (274 patients) No sequelae Major sequelae
79 (29) 19 (7)
XRT and surgery (29 patients) No sequelae Major sequelae
5 (17) 5 (17)
Salvage surgery (1 O0 patients) No sequelae Major sequelae
5 (5) 35 (35)
XRT = radiotherapy.
patients (8%) in the esophagus, and 20 patients (5%) in another area, with a mean time to occurrence of 36 months (range: 6 to 184 months). The 5-year actuarial (Table HI) survival was 33% for the overall population, and 42% for patients who underwent operation first (with postoperative XRT); this may be compared with 16% for patients who underwent operation after preoperative XRT, or with 17% for patients who underwent salvage surgery. Of course, there was no reason to statistically compare prior surgery with salvage surgery, but comparison between surgery with postoperative XRT and surgery with preoperative XRT showed a statistical advantage for postoperative XRT (p = 0.01). "Direct failures" that were truly related to the therapeutic protocol (that is, postoperative deaths and locoregional failures) (Table IV) occurred in 31% of patients with postoperative XRT, compared with 59% of patients who had preoperative XRT (p = 0.03). In the case of salvage surgery, these failures occurred in 61% of patients. In other words, salvage surgery was successful in 39% of patients. Whatever the protocol, local and/or regional failures were rarely controlled, since 66 of 88 patients (75%) with postoperative XRT, 16 of 17 patients (94%) with preoperative 438
THE AMERICAN JOURNAL OF SURGERY
XRT, and 46 of 59 patients (78%) with salvage surgery died of this evolution. Surprisingly, there was no difference, in this regard, between the group of patients who underwent operation as a first attempt and the group of patients who underwent a salvage therapy operation (p = 0.8). Functional results: Due to the segmental mandibulectomy and, as the case arose, the base-of-tongue resection, the functional results (and, subsequently, quality of life) were rather disappointing for patients. Ninety-one patients (23%) were able to resume normal oral feeding after treatment, whereas 245 patients (61%) complained of having to follow a blenderized diet regimen and 59 patients (15%) of having to follow a soft diet (3 even had a gastrostomy). Two hundred six patients (61%) were able to maintain normal speech, 131 patients (32%) were left with a minor speech impairment, and 7 were left with a major impairment. Of course, negative sequelae were more frequent in the case of salvage surgery (Table V), which was associated with larger resections. COMMENTS AND CONCLUSION In our experience, surgery as a first therapeutic choice is suitable for resectable tumors of the PLOC or LO involving the mandible and/or the base of the tongue without reaching the midline. Postoperative courses were acceptable, with a postoperative death rate of 2% and a mean postoperative stay of 15 days. Locoregional failures occurred in 35% of patients and remained uncontrolled in 22% of patients. The 5-year survival was 42%. In our experience, preoperative XRT worsened results, and, in a soon-to-be published randomized study by the French Head and Neck Group (GETTEC), induction chemotherapy with three cycles of cisplatin-5fluorouracil prior to standard treatment (definitive XRT or surgery with postoperative XRT according to institutional protocols) without modification of treatment, whatever the response to chemotherapy, did not seem, in preliminary results, to have a significant impact on the therapeutic outcome. Salvage surgery must be performed each time a persistent or a recurrent disease after XRT is resectable. Postoperative courses were acceptable, with a postoperative death rate of 2% and a postoperative stay of 21 days. New local and/or regional failures occurred in 59% of patients, and remained uncontrolled in 46% of patients. The 5-year survival after surgery was 17%. The use of the pectoralis major myocutaneous flap appears worthwhile and has definitively modified the postoperative courses of salvage surgery with the near-total nonoccurrence of fistulas and hemorrhages. Nevertheless, if cosmetic results are quite acceptable in most cases, functional results are more questionable. The use of dental protheses is feasible and efficient in a very limited number of patients. Segmental mandibulectomy and base-of-tongue resection have a negative impact on chewing and swallowing functions in 77% of the patients, resulting in undernutrition and poor quality-oflife in 15% of patients. With this in mind, we are modifying our present therapeutic protocols. The mandible will
VOLUME 166
OCTOBER 1993
COMPOSITE RESECTION IN BUCCO-PHARYNX CANCER
be preserved when possible as a result of the "mandibular swing" technique for tumors remaining at a distance from the mandible or by the successful use of marginal resection. We are initiating a program of hemi-mandibular reconstruction using, possibly, a titanium prosthesis with auto- or allogenic bone or, preferably, free bone transfers (iliac crest or fibula); such techniques are worthwhile for anterior mandible reconstruction and appear of interest for hemi-mandible reconstruction, as well. Techniques for soft tissue repair should be selected according to what tissue requires reconstruction. For example, the use of the pectoralis major myocutaneous flap
remains the simplest safe technique for a "volume reconstruction," whereas free transfers (gastric or jejunal) allow "surface reconstruction/' with notable advantages such as self-lubricating mucosa and indisputable pliability. In addition, the use of sensate free flaps (radial forearm) seems to be a valuable research avenue for pharyngeal reconstruction. Finally, if one has any doubt about the value of prior surgery for tumors involving the mandible, we think that surgery for tumors involving the base of the tongue should be compared with those treated by concomitant chemo-XRT or hyperfractionated XRT.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 166
OCTOBER 1993
439
From 1972 to 1987, 403 patients underwent a composite resection consisting of segmental mandibulectomy and neck dissection for the treatment of bueco-pharynx squamous cell carcinoma (303 patients had postoperative radiotherapy [XRT], 29 patients had preoperative XRT, and 100 patients had tissue salvage performed after XRT). Of the 303 patients who received preoperative XRT, 32 were clinically staged T2, 149 were staged T3, and 122 were staged T4; 194 of the 303 patients were staged NO. In the 100 patients who experienced relapse and who required "salvage surgery," the restaging found 20 patients staged T2, 39 staged T3, and 41 staged T4; 73 of the 100 patients were staged NO. The patients' mean postoperative stay was 15 days for those who had prior surgery and 21 days for those who had salvage surgery. With a minimum follow-up of 5 years, locoregional recurrences and postoperative death occurred in 86 of 2 7 4 patients (31%) in whom surgery and postoperative XRT were performed; in 17 of 29 patients (59%) who had preoperative XRT performed; and in 61 of 100 patients (61%) who had salvage surgery performed. In addition, in terms of functional results, only 61% of patients ( 2 0 6 ) were able to maintain normal speech function, and only 23% ( 9 1 ) were able to maintain normal oral feeding. Of the 403 patients, the 5-year survival rates were 33% for the overall population, 42% for patients with postoperative XRT, 16% for patients in whom operation was performed after preoperative XRT, and 17% for patients who underwent salvage surgery.
ince its description by Hayes Martin in the 1940s, resection with segmental mandibulectomy Sand composite en-bloc radical neck dissection has been the procedure of choice in the surgical management of cancer of the posterolateral part of the oral cavity (PLOC) or of the lateral oropharynx (LO). In some cases, either due to mandible invasion or to the need for a proper surgical access for tumor resection, ablation of the posterior part of the mandible (including the ramus) is required. Such surgery has undoubtedly improved local control, but with the subsequent addition of some morphologie and/or functional sequelae. At Centre Oscar Lambret (Northern France Comprehensive Cancer Center), we often encounter buccopharyngeal tumors; approximately 200 patients with squamous cell carcinoma of the oral cavity, the oral tongue, or the oropharynx are referred to our institution yearly. In about half of these patients, the tumor is located in the PLOC or LO. According to our therapeutic protocols, tumors of the retromolar trigone, tumors of the very posterior part of the floor of the mouth (reaching or invading the mandible and for which a marginal mandibular resection would not be sufficient), tumors of the junctional area of the tongue (junction between the anterior pillar and mobile tongue), tumors of the tonsillar fossa (invading pterygoid muscles or the ramus or the glossotonsillar sulcus), and tumors of the glossotonsillar sulcus, itself, are treated by composite resection and postoperative radiotherapy. Other tumors of the PLOC or LO are treated by definitive radiotherapy (with neo-adjuvant chemotherapy for advanced stages), with surgery in reserve for salvage, if necessary. Our experience over a 15-year period is reviewed for this presentation.
PATIENTS AND METHODS Population: From 1972 through 1987, 403 patients were successively treated by surgery consisting of composite resection of the buccopharynx with segmental mandibulectomy and neck dissection (the so-called "commando" operation). There were 376 men with a mean age of 53 years (range: 31 to 82 years) and 27 women with a mean age of From the Departmentsof Head and Neck Surgery (J-LL, BV, EB, 59 years (range: 35 to 83 years). In 190 patients, the JTV), Radiotherapy(BP, BC-D), and Anesthesiology(TO), Centre primary site was the PLOC, that is, the retromolar triOscar Lambret,LilleC&lex,France. gone, the junctional area of the tongue, and the posterior Requests for reprints shouldbe addressedto Jean-LouisLefebre, part of the floor of the mouth. In 213 patients, the priMD, CentreOscar Lambret,Northern FranceComprehensiveCancer mary site was the LO, that is, the tonsillar fossa and Center, B.P. 307, 59020 LilleC&lex,France. Presentedat the 39th AnnualMeetingof the Societyof Head and pillars, and the glossotonsillar sulcus. Neck Surgeons,Los Angeles,California,March 18-21, 1993. Three hundred three patients (75%) were previously THE AMERICAN JOURNAL OF SURGERY VOLUME166 OCTOBER1993 435
LEFEBVRE ET AL
TABLE I
TABLE II
Results of Head and Neck Cancer Staging Criteria* Applied to Patients Who Received Preoperative Radiotherapy Stage of Nodal Involvement
T1
T2
T3
NO N1 N2a N2b N2c N3
----. --
23 5 4 --
101 26 18 3
Total no, of Patients
--
.
Total No: of Patients
Stage of Nodal Involvement
T1
T2
T3
T4
194 57 33 10
NO N1 N2a N2b N2c N3
I ---. --
i7 3 ---
29 5 4 -1
27 4 6 1 . 3
73 12 10 1
1
70 26 11 7 . 8
149
122
303
Total no. of Patients
--
39
41
100
. --
32
Results of Head and Neck Cancer Staging Criteria* Applied to Patients Who Experienced Relapse and Who Were Re.Staged for Salvage Surgery
T4
.
9
*American Joint Committee on Cancer (AJCC) criteria (1988).
.
. --
20
.
Total No. of Patients
4
*American Joint Committee on Cancer (AJCC) Criteria (1988).
/ f untreated. Surgery was the first therapeutic procedure, followed by postoperative radiotherapy (XRT) in 274 patients, whereas, in 29 patients, surgery was performed after preoperative XRT. In contrast, 100 patients (25%) had had previous treatment consisting of full-dose XRT and underwent operation(s) for local (with or without regional) recurrence. Clinical staging was evaluated according to the criteria recommended by the American Joint Committee for Cancer (AJCC, 1988). For accurate diagnosis, the staging was made at the time of decision-making for previously untreated patients. For patients who underwent operation(s) for a local relapse, this re-evaluation was made at the time of surgery. Using this classification, we found that for the 303 previously untreated patients (Table I), 271 patients (89%) had T3 and T4 lesions, and 109 (36%) had neck nodal involvement staged N1 to N3; 23 cancers were classified as stage II, 132 as stage III, and 148 as stage IV. For the restaging of the patients who experienced recurrence (Table II), we found that 80 patients (85%) had T3 and T4 lesions, and 26 patients (26%) had neck 436
THE AMERICAN JOURNAL OF SURGERY
Figure
1. C l a s s i c a l lip-and-chin-splitting i n c i s i o n used a s p a r t o f c o m p o s ite r e s e c t i o n w i t h m a n d i b u l e c t o m y .
nodal involvement staged N1 to N3; 17 cancers were classified as stage II, 37 as stage III, and 46 as stage IV. Surgical proeedutres: Before choosing to go ahead with the operation, all of the patients received a complete evaluation, including the following: panendoscopy, chest radiography, orthopantography, and classical preoperative biologic assessment; all had histologic confirmation of their disease, or of its recurrence (all were confirmed as cases of squamous cell carcinoma). When necessary, patients underwent nutritional support prior to surgery. Dental care was performed, as well. The evaluation of the tumor volume was assessed under general anesthesia (during panendoscopy), most often through a joint examination by a head and neck surgeon and a radiation specialist. Surgery was performed under general anesthesia with tracheotomy. We used a classical lip-and-chin-splitting incision (Figure 1), which recently has been replaced by a lip-splitting~-chin-sparing incision. The neck dissection was, in all cases, the first surgical procedure. In the case of neck nodal involvement staged NO or N1, we performed a modified dissection. In the case of palpable
VOLUME 166
OCTOBER 1993
COMPOSITE RESECTION IN BUCCO-PHARYNX CANCER
lymph nodal tumors over 3 cm in diameter (N2), a traditional radical neck dissection was performed; the spinal accessory nerve was preserved, whenever possible. Thereafter, the tumor excision was initiated: the mandible was sawed at the level of the mental foramen and removed after exarticulation of the temporomandibular joint, in continuity with the tumor. Surgical margins were assessed by frozen sections. During this period of time (1972 to 1987), we did not perform any mandibular repair operations. In contrast, we used, in 43 patients (11%), a pectoralis major myocutaneous flap for soft tissue repair. All previously untreated patients underwent a neck dissection, whereas 63 of the patients who experienced a recurrence underwent some surgery in the neck. The 37 patients who did not undergo neck dissection either had had prior surgery in the neck or had residual sclerosis without palpable lymph nodal tumors after XRT; these factors disinclined us to perform a neck dissection in these patients. Radiotherapy protocol: Patients who were treated for a local recurrence of cancer had received definitive XRT that delivered 60 to 65 Gy (10 Gy per week, 1 dose of 2 Gy per day) up to 70 to 75 Gy on residual disease or palpable lymph node involvement, using two opposite portals. Sources were a cobalt 60 machine and an electron beam for boosting. At 40 Gy, the field was reduced in order to spare the spine. Patients underwent operation at the time of recurrence or, in the case of persistent disease, 3 months after the completion of XRT. Patients who had preoperative XRT received 45 Gy over a 3-week period (15 Gy per week, 1 dose of 3 Gy per day) with similar fields, by means of a cobalt 60 machine. These patients underwent operation after a 2-week rest. Fifteen days after wound healing, patients who had postoperative XRT received 55 Gy on both the neck and pharynx--up to 70 Gy on positive margins or extracapsular spread areas; cobalt 60 and electron beam were used, which is similar to the protocol followed for definitive XRT. Finally, patients who underwent salvage surgery with myocutaneous flap received depilation XRT (if necessary) of 30 Gy electron beam (10 Gy per day) in the pectoralis area, over the 3 days preceding surgery. Statistical considerations: Survival rates were calculated according to the Kaplan-Meier method, and survival rates were compared by the log-rank test. Survival was evaluated from the day of pathologic diagnosis for previously untreated patients, and from the day of surgery for those patients who underwent salvage surgery. Comparison of effectiveness was statisticaUy analyzed with the x 2 test. RESULTS
Postoperative course: Eight patients (2%) died during the postoperative course: two of severe dyspnea (patients who had no tracheotomy but whose dyspnea very quickly became systematic), three of myocardiac infarction, two of multi-system organ failure, and one of massive aspiration. Thirty-one patients (8%) had a delay in wound healing. The mean tracheotomy duration was 12
TABLE
III
Actuarial Survival Rates by Percent for Patients at 3-Year, 5-Year, and 10-Year Follow-Up
Patient Groups
Percent of Percent of Percent of Total No. Patients Patients Patients of Patients at 3 Years at 5 Years at 10 Years
Overall population
403
48
33
17
Surgery and postoperative XRT
274
55
42
21
Preoperative XRT and surgery
29
24
16
16
Salvage surgery
100
23
17
NE
XRT = radiotherapy
days (range: 2 to 55 days), and two patients had delayed removal of the tracheotomy cannula (more than 3 weeks). The mean length of time the nasogastric feeding tube was in place was 13 days (range: 4 to 90 days); 22 (5%) patients had delayed removal of the feeding tube (more than 3 weeks), of whom 3 required a gastrostomy. For the overall population, the postoperative in-hospital stay was 17 days (range: 8 to 90 days): 15 days for patients who underwent prior surgery and 21 days for patients who underwent salvage surgery. Pathology: Frozen section revealed satisfactory results at first resection in 70% of patients; 25% needed further resection; and 4% were not evaluable, due to fragmentation of the specimen. On paraffin section analysis, surgical margins were negative in 76% of patients, suspicious in 17%, positive in 6%, and not evaluable in 1% (fragmented specimen). In neck dissection, there were 168 patients (46%) without nodal involvement, 80 patients (23%) with nodal involvement but without extracapsular spread, and 105 patients (29%) with extracapsular spread. In 111 patients (30%), there were 1 or 2 involved lymph node(s), and, in 80 patients (22%), 3 or more nodes were involved. Finally, in seven patients (2%) who underwent operation after definitive XRT, it was impossible to assess whether the lymph nodes were still evolutive or not. Long-term results: To date, with a minimum followup of 5 years, 133 patients (33%) have not demonstrated any cancer evolution nor second primary-site appearance. One hundred sixty patients (40%) demonstrated a cancer evolution but no second primary site, 86 patients (21%) demonstrated 1 second primary site, and 16 patients (4%) demonstrated both a cancer evolution and the presence of a second primary site. Local relapse occurred in 122 patients (30%) and nodal recurrence in 48 patients (12%). Both events occurred very quickly, with a mean time to occurrence of 9 months (range: 6 to 60 months, 98% within 24 months). Distant metastasis occurred in 38 patients (9%). Seventy-one (18%) patients developed a second primary site in the upper aerodigestive tract, 31
THE AMERICAN JOURNAL OF SURGERY
VOLUME 166
OCTOBER 1993
437
LEFEBVREET AL
TABLE IV " D i r e c t F a i l u r e s " Related to Therapeutic Protocol at a
Minimum Follow-Up of 5 Years
No. of Cases of Postoperative Death
Cases of Local and/or Regional Recurrence (*)
Totals (%)
Surgery and postoperative XRT
6
80 (66)
31
Preoperative XRT and surgery
--
17 (16)
59
2
59 (46)
61
Therapeutic Protocol Followed
Salvage surgery
*Number of patients who died due to uncontrolled local and/or regional recurrenceof cancer. XRT = radiotherapy,
TABLE V
Protocol Followed and Incidence of Negative Sequelae--Functional Results Protocol Followed and Patient Groups
No. of Patients (%)
Surgery and XRT (274 patients) No sequelae Major sequelae
79 (29) 19 (7)
XRT and surgery (29 patients) No sequelae Major sequelae
5 (17) 5 (17)
Salvage surgery (1 O0 patients) No sequelae Major sequelae
5 (5) 35 (35)
XRT = radiotherapy.
patients (8%) in the esophagus, and 20 patients (5%) in another area, with a mean time to occurrence of 36 months (range: 6 to 184 months). The 5-year actuarial (Table HI) survival was 33% for the overall population, and 42% for patients who underwent operation first (with postoperative XRT); this may be compared with 16% for patients who underwent operation after preoperative XRT, or with 17% for patients who underwent salvage surgery. Of course, there was no reason to statistically compare prior surgery with salvage surgery, but comparison between surgery with postoperative XRT and surgery with preoperative XRT showed a statistical advantage for postoperative XRT (p = 0.01). "Direct failures" that were truly related to the therapeutic protocol (that is, postoperative deaths and locoregional failures) (Table IV) occurred in 31% of patients with postoperative XRT, compared with 59% of patients who had preoperative XRT (p = 0.03). In the case of salvage surgery, these failures occurred in 61% of patients. In other words, salvage surgery was successful in 39% of patients. Whatever the protocol, local and/or regional failures were rarely controlled, since 66 of 88 patients (75%) with postoperative XRT, 16 of 17 patients (94%) with preoperative 438
THE AMERICAN JOURNAL OF SURGERY
XRT, and 46 of 59 patients (78%) with salvage surgery died of this evolution. Surprisingly, there was no difference, in this regard, between the group of patients who underwent operation as a first attempt and the group of patients who underwent a salvage therapy operation (p = 0.8). Functional results: Due to the segmental mandibulectomy and, as the case arose, the base-of-tongue resection, the functional results (and, subsequently, quality of life) were rather disappointing for patients. Ninety-one patients (23%) were able to resume normal oral feeding after treatment, whereas 245 patients (61%) complained of having to follow a blenderized diet regimen and 59 patients (15%) of having to follow a soft diet (3 even had a gastrostomy). Two hundred six patients (61%) were able to maintain normal speech, 131 patients (32%) were left with a minor speech impairment, and 7 were left with a major impairment. Of course, negative sequelae were more frequent in the case of salvage surgery (Table V), which was associated with larger resections. COMMENTS AND CONCLUSION In our experience, surgery as a first therapeutic choice is suitable for resectable tumors of the PLOC or LO involving the mandible and/or the base of the tongue without reaching the midline. Postoperative courses were acceptable, with a postoperative death rate of 2% and a mean postoperative stay of 15 days. Locoregional failures occurred in 35% of patients and remained uncontrolled in 22% of patients. The 5-year survival was 42%. In our experience, preoperative XRT worsened results, and, in a soon-to-be published randomized study by the French Head and Neck Group (GETTEC), induction chemotherapy with three cycles of cisplatin-5fluorouracil prior to standard treatment (definitive XRT or surgery with postoperative XRT according to institutional protocols) without modification of treatment, whatever the response to chemotherapy, did not seem, in preliminary results, to have a significant impact on the therapeutic outcome. Salvage surgery must be performed each time a persistent or a recurrent disease after XRT is resectable. Postoperative courses were acceptable, with a postoperative death rate of 2% and a postoperative stay of 21 days. New local and/or regional failures occurred in 59% of patients, and remained uncontrolled in 46% of patients. The 5-year survival after surgery was 17%. The use of the pectoralis major myocutaneous flap appears worthwhile and has definitively modified the postoperative courses of salvage surgery with the near-total nonoccurrence of fistulas and hemorrhages. Nevertheless, if cosmetic results are quite acceptable in most cases, functional results are more questionable. The use of dental protheses is feasible and efficient in a very limited number of patients. Segmental mandibulectomy and base-of-tongue resection have a negative impact on chewing and swallowing functions in 77% of the patients, resulting in undernutrition and poor quality-oflife in 15% of patients. With this in mind, we are modifying our present therapeutic protocols. The mandible will
VOLUME 166
OCTOBER 1993
COMPOSITE RESECTION IN BUCCO-PHARYNX CANCER
be preserved when possible as a result of the "mandibular swing" technique for tumors remaining at a distance from the mandible or by the successful use of marginal resection. We are initiating a program of hemi-mandibular reconstruction using, possibly, a titanium prosthesis with auto- or allogenic bone or, preferably, free bone transfers (iliac crest or fibula); such techniques are worthwhile for anterior mandible reconstruction and appear of interest for hemi-mandible reconstruction, as well. Techniques for soft tissue repair should be selected according to what tissue requires reconstruction. For example, the use of the pectoralis major myocutaneous flap
remains the simplest safe technique for a "volume reconstruction," whereas free transfers (gastric or jejunal) allow "surface reconstruction/' with notable advantages such as self-lubricating mucosa and indisputable pliability. In addition, the use of sensate free flaps (radial forearm) seems to be a valuable research avenue for pharyngeal reconstruction. Finally, if one has any doubt about the value of prior surgery for tumors involving the mandible, we think that surgery for tumors involving the base of the tongue should be compared with those treated by concomitant chemo-XRT or hyperfractionated XRT.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 166
OCTOBER 1993
439