T vancements in the diagnosis and treatment of cancer.

Transcription

1 Esophageal Carcinoma: Esophageal Ultrasound ssessment of Preoperative Chemotherapy Thomas W. Rice, MD, Gregory. oyce, MD, Michael V. Sivak, MD, David J. delstein, MD, and Thomas J. Kirby, MD Departments of Thoracic and Cardiovascular Surgery, Gastroenterology, and Hematology and Oncology, The Cleveland Clinic Foundation, Cleveland, Ohio, and Department of Gastroenterology, University of South Florida, Tampa, Florida The effect of preoperative chemotherapy in the treatment of esophageal carcinoma is difficult to assess because of the inadequacies of clinical staging. Endoscopic esophageal ultrasound (EUS) has been shown to be accurate in the clinical determination of depth of tumor invasion (T) and regional lymph node status (N). Therefore, EUS may be useful in assessing the effect of preoperative chemotherapy in the treatment of esophageal carcinoma. Eleven patients with operable adenocarcinoma of the esophagus or esophagogastric junction underwent staging by EUS before treatment. This was followed by two courses (10 patients) or one course (1 patient) of chemotherapy: etoposide, 120 mg/mz for 3 days; doxorubicin hydrochloride, 20 mg/m2; and cisplatin, 100 mg/m'. Restaging by EUS was done after treatment. Ten patients then underwent resection of the tumor with lymphadenectomy. One patient was found to have metastatic disease at thoracotomy and did not undergo resection. However, tissue sampling was adequate for the determination of pathological stage. Independent pathological determinations of T and N were then obtained. On completion of chemotherapy, 9 patients (82%) had relief or reduction of preoperative symptoms, and 9 patients (82%) had either no evidence of tumor or reduction of tumor size by endoscopy. Despite this clinical and endoscopic response, no patient had EUS-documented and pathologyconfirmed reduction of T. However, 2 patients had EUSdocumented and pathology-confirmed progression of N. The accuracy of EUS in the determination of T was 82% and of N, 73%. We conclude that EUS offers a unique evaluation of the esophageal wall and adjacent tissue and is superior to symptomatic or endoscopic assessment of the effect of preoperative chemotherapy in the treatment of esophageal carcinoma. (nn Thoruc Surg 2992;53:972-7) he staging of malignancies has improved with ad- T vancements in the diagnosis and treatment of cancer. The stage of a tumor can be determined by diagnostic evaluation before treatment (clinical staging) or by evaluation of the resected specimen (pathological staging). ccurate pathological staging is crucial for the prognosis, management, and study of patients after resection. Clinical staging, although less precise than pathological staging, is critical in the selection of appropriate treatment [l]. In addition, clinical staging provides an objective means for assessing the response to various agents in controlled trials of multimodality therapy [Z, 31. Previous studies [P7] have used relief of preoperative symptoms, barium esophagram, esophagoscopy, or computed tomographic scan (CT) to assess the clinical stage before and after treatment and to evaluate the degree of response of esophageal carcinoma to preoperative therapy. However, the accuracy of these clinical staging techniques in predicting the pathological stage is not optimal. The recent introduction of endoscopic esophageal ultrasound (EUS), an advancement in the staging of esopha- Presented at the Thirty-eighth nnual Meeting of the Southern Thoracic Surgical ssociation, Orlando, FL, Nov 7-9, ddress reprint requests to Dr Rice, The Cleveland Clinic Foundation, 9500 Euclid ve, Cleveland, OH geal carcinoma, has prompted the introduction of a new staging system for esophageal carcinoma. This current staging system was developed by the merican Joint Committee on Cancer (JCC) [8] in conjunction with the International Union gainst Cancer [9] and uses the TNM format. T characterizes the primary tumor by measuring the depth of invasion of the esophageal wall by the primary tumor; N characterizes the regional lymph nodes by assessing them for the presence of metastases; and M characterizes sites of potential distant metastatic disease. The TNM sets are then arranged into staging groups that rank tumors from those with the best prognosis (stage 0) to those advanced tumors with the worst prognosis (stage IV). The staging is as follows: T: Primary Tumor TX Primary tumor cannot be assessed TO No tumor Tis Carcinoma in situ T1 Tumor invades the lamina propria or submucosa T2 Tumor invades the muscularis propria T3 Tumor invades the periesophageal tissue T4 Tumor invades adjacent structures N: Regional Lymph Nodes NX Regional nodes cannot be assessed NO No regional lymph node metastases N1 Regional lymph node metastases by The Society of Thoracic Surgeons /92/$5.00

2 nn Thorac Surg 1992;53:972-7 RICEETL 973 EUS SSESSMENT OF PREOPERTIVE CHETHERPY M: Distant Metastases MX Distant metastases cannot be assessed No distant metastases M1 Distant metastases Staging Groups Stage 0 Tis NO Stage I T1 NO Stage IIa T2 NO T3 NO Stage IIb T1 N1 T2 N1 Stage 111 T3 N1 T4 ny N Stage IV ny T ny N M1 Esophageal ultrasound enables the endoscopic placement of a high-frequency ultrasound transducer in close proximity to the esophageal wall and periesophageal tissues. The esophageal wall is imaged as a five-layered structure by EUS, allowing depth of tumor invasion (T) to be assessed. This technique also provides a unique visualization of regional lymph nodes (N). Size, shape, margin, and echo characteristics can be evaluated. The celiac axis lymph nodes, now considered a site of distant metastases (M), can also be visualized. Several recent prospective studies [1CL13] have demonstrated that clinical determination of T and N of esophageal cancer with EUS is superior to CT. Therefore, EUS may be a more accurate means of assessing the therapeutic effect of preoperative chemotherapy in the treatment of esophageal carcinoma. Material and Methods Thirteen patients with carcinoma of the esophagus constituted the study cohort, and 11 completed the study. ll patients were men. ge ranged from 50 to 69 years (median age, 60 years; mean age, 60.7 years). The presence of biopsy-proven adenocarcinoma of the esophagus or esophagogastric junction with no clinical evidence of distant metastatic disease (beyond the celiac axis), medical suitability for esophageal resection, and an Eastern Coop- erative Oncology Group performance status of 0 to 2 were the eligibility criteria. Patients who had received any previous therapy for the malignancy or who had any evidence of major preexisting renal or hematological dysfunction were excluded. ll patients were assessed by a thoracic surgeon, a gastroenterologist, and a medical oncologist. This study was approved by the Cleveland Clinic Foundation Research Programs Committee (November 21, 1987), and appropriate informed consent was obtained before treatment was begun. Pretreatment staging evaluation included a medical history, physical examination, complete blood count, electrocardiogram, urinalysis, and serum chemistry tests including urea nitrogen, creatinine, calcium, phosphorus, alkaline phosphatase, serum glutamic-oxaloacetic transaminase, lactate dehydrogenase, albumin, total protein, bilirubin, and uric acid. ll patients underwent baseline barium esophagraphy, chest radiography, CT of the chest and abdomen, and pulmonary function studies. ronchoscopy was performed only when indicated by the extent or location of the primary lesion. Esophagogastroduodenoscopy, endoscopic esophageal biopsy, and endoscopic EUS were performed in all patients. Patients were premedicated with meperidine hydrochloride, 50 mg, and midazolam hydrochloride, 2 mg, intravenously. Local anesthesia was not required. During the upper gastrointestinal endoscopy study, biopsy specimens of the tumor were obtained for pathological confirmation in all 13 patients (8 had arrett mucosa). n Olympus EU-M3 ultrasound endoscope (Olympus Corporation, Lake Success, NY) was passed blindly and positioned endoscopically to examine the entire esophagus, upper portion of the stomach, adjacent segment of the liver, and celiac axis. The examinations were videotaped. Using the JCC staging system, the pretreatment clinical determination of T and N was recorded [14]. When possible, adjacent sites of metastases were studied, and an attempt was made to define the status of distant metastatic sites (M) by EUS. Patients with T14 NO-1-1 (M1 = celiac lymph node metastases only) disease were eligible for preoperative therapy. Fig I. () hepatic metastasis (arrow) in the left lateral segment of the liver. The esophageal ultrasound probe is seen in the gastric cardia. () The hepatic metastasis (arrow) as seen from the gastric cardia by esophageal ultrasound. The metastasis was imaged only by esophageal ultrasound.

3 974 RICEETL EUS SSESSMENT OF PREOPERTIVE CHETHERPY nn Thorac Surg 1992; Fig 2. () The hepatic metastasis (upper arrow) has increased in size after two courses of chemotherapy. regional lymph node metastasis (lower arrow) has developed. () The hepatic metastasis (upper arrow) was shown to have increased in size by esophageal ultrasound. The regional lymph node metastasis flower arrow) was seen by esophageal ultrasound. The chemotherapy induction regimen was a modified version of the etoposide, doxorubicin hydrochloride, cisplatin combination used by Wilke and colleagues [15]. Two courses of chemotherapy were planned: etoposide, 120 mg/m on days 1,2, and 3; doxorubicin, 20 mg/m on day 1; and cisplatin, 100 mg/m2 on day 2. The cisplatin administration necessitated hospitalization and was accompanied by vigorous intravenous hydration and diuresis induced by furosemide and mannitol. ggressive antiemetic management including metoclopramide and dexamethasone was provided for all patients. The chemotherapy was repeated in 4 weeks for all patients except 1, who experienced sufficient toxicity that only a single Chemotherapy course was given. No dosage modifications for myelosuppression were made because only two courses of chemotherapy were planned. The second course of chemotherapy could be delayed for 1 week to allow for full hematological recovery, and the cisplatin dose could be reduced or deleted if any persistent renal functional impairment or hearing loss developed. ll patients were followed closely during their induction chemotherapy in an effort to detect and manage any treatment-induced toxicities. Four weeks after the final chemotherapy, patients underwent clinical assessment of symptomatic improvement, restaging esophagogastroduodenoscopy and EUS, chest radiography, and blood chemistry studies. Computed tomography and other staging procedures were repeated only if clinically indicated. fter chemotherapy, an appraisal of tumor response based on symptom assessment, gross endoscopic appearance, and an EUSdetermined clinical stage was assigned to each patient at the time of this reevaluation. If there was no evidence of distant metastatic disease or unresectability, the patient underwent surgical intervention. Two patients were found to have distant metastatic disease at exploratory laparotomy and thoracotomy. They did not undergo resection. However, there was adequate tissue sampling to determine the pathological stage of the disease in 1 of these patients. The remaining 11 patients underwent a left thoracoabdominal incision for resection of the primary tumor and lymphadenectomy. In 6 patients, a left neck incision placed obliquely along the border of the left sternocleidomastoid muscle was used to complete the esophagectomy and allow reconstruction. total gastrectomy and distal esophagectomy was required in 1 patient. Examination of the resected esophagus and periesophageal tissue was performed, and a pathological stage was assigned. Clinical stage before and after chemotherapy, gross tumor response, and symptomatic response were then compared and correlated with the pathological stage. Fig 3. () Esophagoscopy of an esophageal carcinoma before chemotherapy. The bulky tumor nearly obstructed the lumen, but the esophageal ultrasound probe could be passed through the tumor. () fter two courses of Chemotherapy, there was a reduction of tumor bulk (partial response).

4 nn Thorac Surg 1992; RICE ET L EUS SSESSMENT OF PREOPERTIVE CHETHERPY 975 Table 1. Clinical and Pathological Stages of Esophageal Carcinoma Patient No. Clinical Pretreatment Posttreatment Pathological T N M Stage T N M Stage T N M Stage IV IV IV IIa I I IIa I IIa IIa IIa I I IIa IIa IIb I I I I IV IIa IIa IIb IIa Results Two patients (15%) did not complete the study. One of them was found to have a solitary hepatic metastasis in the left lateral segment of the liver (Figs 1, 2). This metastasis was not seen by any imaging modality except EUS, and its presence was confirmed by exploratory laparotomy and biopsy. The other patient did not have pretreatment staging because of the inability to pass the ultrasound endoscope through the malignant esophageal stricture. lthough this patient was excluded from the study, it was possible to stage the tumor by EUS after the completion of two courses of chemotherapy. Eleven patients (85%) completed the study. Nine (82%) patients, 8 with dysphagia and 1 with odynophagia, had reduction or relief of the preoperative,symptoms. The tumor mass was reduced at endoscopy in 9 patients (82%), with one complete response (CR) and eight partial responses (PRs) (Fig 3). The pretreatment and posttreatment clinical stages and the pathological stages of the 11 patients are shown in Table 1. Depth of tumor invasion (T) was correctly predicted in 9 of 11 patients by EUS, an 82% accuracy rate. One patient was understaged, and 1 patient was overstaged, each by one level. The predictive power of T3 by EUS was 88% (7/8); the predictive power of T2 by EUS was 67% (2/3). In all but 1 patient, EUS demonstrated no reduction in T after chemotherapy. The one EUS study that predicted reduction was the result of understaging in the posttreatment evaluation, as the T status remained unchanged from pretreatment to resection. Regional lymph node status (N) was correctly predicted in 8 of 11 patients by EUS, a 73% accuracy rate. There were three incorrect clinical assessments of N; all were false-negative. The sensitivity for N was 70% and the specificity, 100%; the positive predictive value was 100% and the negative predictive value, 25%. In 1 patient, a reduction of N predicted by EUS was not seen in the pathological review of the resected specimen. In 2 patients (18%), N progression predicted by EUS was confirmed by pathological evaluation (Figs 4, 5). There were three instances of metastases to distant sites among the 13 patients, one each to lung, liver, and celiac lymph nodes. Esophageal ultrasound detected the metastasis in the liver and celiac lymph nodes because these sites were in direct contact with the stomach during the examination. This technique could not detect the pulmonary metastasis because of the intervening air-filled lung, which did not allow an acceptable acoustic interface between the esophagus and the metastasis. The stage of disease was correctly identified in 7 (64%) of 11 patients by EUS. ll errors were the result of understaging, three by one stage (two stage IIb tumors misinterpreted as stage IIa tumors and one stage IV tumor misinterpreted as a stage I11 tumor) and one by two stages (a stage I11 tumor misinterpreted as a stage IIa tumor). One of the errors was the result of the inability of EUS to detect the pulmonary metastasis. The other three errors were due to EUS misinterpretation of N. In 1 patient, the combination of an understaging of T by one level and the false-negative interpretation of N resulted in the understaging of the tumor by two levels. The other incorrect determination of T did not affect the overall stage, as there were no regional lymph node metastases, and the stage remained IIa. Comment The prediction of pathological stage by clinical assessment is crucial if the effectiveness of preoperative chemotherapy in the treatment of esophageal carcinoma is to be Fig 4. () T3 NO tumor before chemotherapy. () The T3 NO tumor as seen by esophageal ultrasound. (0 = aorta.)

5 976 RICEETL EUS SSESSMENT OF PREOPERTIVE CHETHERPY nn Thorac Surg 1992;53:972-7 Fig 5. () The T3 NO tumor in Figure 4 after two courses of chemotherapy. regional lymph node metastasis (arrow) has developed, and the tumor is no% T3 N1. () The progression of the tumor to T3 Nl was documented by esophageal ultrasound, and the regional lymph node metastasis (arrow) was visualized. (0 = aorta.) evaluated. Conventional staging tools such as esophagoscopy and barium esophagram demonstrate only the intraluminal extent of disease. Computed tomography of the chest is relatively insensitive for clinical staging of esophageal carcinoma except for extensive local disease and distant metastases. ssessment of symptoms is subjective and of limited value in the determination of clinical stage. None of these conventionally employed staging methods is adequate in the assessment of either the depth of esophageal wall invasion (T) or the presence of regional lymph node involvement (N). It is now known that these measures of malignant disease are of great prognostic importance in esophageal and gastrointestinal malignancies [I, 8, 91. In 1983, the JCC proposed that a clinical stage be defined according to the length of intraluminal esophageal tumor, the percentage of luminal obstruction, and the presence of palpable (cervical and supraclavicular) lymph nodes. This clinical staging system proved to be of limited value. The 1988 JCC revision now recognizes only a pathological staging system based on the depth of esophageal wall invasion (T) and the presence of regional lymph node (N) and distant (M) metastases. Therefore, any approach to clinical staging must address these same variables. This necessitates the development of noninvasive clinical staging techniques that offer a fair prediction of pathological stage. The availability of such methods would aid in the selection of treatment modalities, allow study of these modalities, and afford accurate prognostication. With the increasing use of preoperative therapy, either induction chemotherapy or neoadjuvant chemotherapy and radiation therapy, the need for such a clinical staging method has increased. Induction chemotherapy studies have failed to accurately assess the effects of preoperative chemotherapy and to correlate a clinical response with surgical findings. Forastiere and co-workers [4] reported a 47% response rate (1 complete response [CR] and 15 partial responses [PRs]) with cisplatin, vinblastine, and mitoguazone induction chemotherapy. The response rate was defined by reduction of tumor length as determined by barium esophagram. One CR was confirmed at operation. For patients with a PR, the only correlation to the surgical findings was that posttreatment tumor lengths were within 1 cm of measured tumor lengths at resection in 80% of patients. Kies and associates [5] reported a response rate of 42% (3 CRs and 8 PRs) in patients treated with cisplatin and 5-fluorouracil. CR was defined as elimination of dysphagia with no evidence of tumor by esophagoscopy or esophagogram. PR was defined as reduction of dysphagia and reduction of tumor size by esophagoscopy, esophagram, or CT. The difficulty in determining a PR by standard means was noted. In addition, the staging after induction chemotherapy failed to detect mediastinal or abdominal metastases in 9 of 14 patients having operation. Residual tumor was found in all resected specimens. There was no further correlation or validation of the clinical responses with the pathological findings. Hilgenberg and colleagues [6] had a response rate of 57% (13 CRs and 7 PRs) in patients treated preoperatively with cisplatin and 5-fluorouracil. response was defined as symptomatic relief and reduction of tumor size by CT, barium esophagram, or esophagoscopy. There was one pathologically confirmed CR, and no further correlation between response to chemotherapy and surgical findings. jani and co-workers [7] obtained a response rate of 49% (six CRs and 11 PRs) in patients receiving etoposide, 5-fluorouracil, and cisplatin induction chemotherapy. Responses were based on barium swallow, endoscopy with biopsy or cytology brushings, and CT. patient with normalization of all studies and negative endoscopic biopsy specimens or cytological results was considered to have a CR. marked reduction in tumor bulk was defined as a PR. On review of the resected specimen, there was only one CR. There was no further correlation between the clinical response and the stage. In this study, there was an exceedingly high response rate (82%) of esophageal carcinoma treated with induction chemotherapy when assessed by symptom relief or endoscopy. These findings were not confirmed by EUS or pathologically determined stage. These results indicate that tumor size and symptom control are poor indicators of clinical stage. The JCC recommends the variables of the TNM system as being the most important factors in the assessment of esophageal carcinoma. ll tumors should, therefore, be assessed according to the TNM system. Currently, EUS is the most accurate method for

6 nn Thorac Surg 1992; RICE ET L EUS SSESSMENT OF PREOPERTIVE CHETHERPY 977 clinical TNM staging. Esophageal ultrasound accurately predicts depth of tumor invasion (T) and regional lymph node status (N) and is superior to symptomatic or endoscopic assessment of the response of esophageal carcinoma to preoperative chemotherapy. References 1. Skinner D, Ferguson MK, Soriano, Little G, Staszak VM. Selection of operation for esophageal carcinoma based on staging. nn Surg 1986;204: Kelsen DP, Heelan R, Coonley C, et al. Clinical and pathological evaluation of response to chemotherapy in patients with esophageal carcinoma. m J Clin Oncol 1983;6: gha FP, Gennis M, Orringer M, Forastiere. Evaluation of response to preoperative chemotherapy in esophageal and gastric cardia cancer using biphasic esophagrams and surgical-pathologic correlation. m J Clin Oncol 1986;9: Forastiere, Gennis M, Orringer M, gha FP. Cisplatin, vinblastine, and mitoguazone chemotherapy for epidermoid and adenocarcinoma of the esophagus. J Clin Oncol 1987;5: Kies MS, Rosen ST, Tsang TK, et al. Cisplatin and 5-fluorouracil in the primary management of squamous esophageal cancer. Cancer 1987;60: Hilgenberg D, Carey RW, Wilkins EW Jr, Choi NC, Mathisen DJ, Grillo HC. Preoperative chemotherapy, surgical resection, and selective postoperative therapy for squamous cell carcinoma of the esophagus. nn Thorac Surg 1988;45: jani J, Roth ], Ryan, et al. Evaluation of pre- and postoperative chemotherapy for resectable adenocarcinoma of the esophagus or gastroesophageal junction. J Clin Oncol 1990;8: eahrs OH, Henson DE, Hutter RVP, Myers MH. Manual for staging of cancer. 3rd ed. Philadelphia: J Lippincott, 1988: Spiessl, eahrs OH, Hermanek P, et al. TNM atlas: illustrated guide to the TNWpTNM classification of malignant tumours. 4th ed. erlin: Springer, 1990:6> Ziegler K, Sanft C, Semsch 8, et al. Endosonography is superior to computed tomography in staging tumors of the esophagus and the cardia [bstract]. Gastroenterology 1988; 94: Tio TL, Cohen P, Coene PP, et al. Endosonography and computed tomography of esophageal carcinoma. Gastroenterology 1989;96: otet JF, Lightdale CJ, Zauber G, Gerdes H, Urmacher C, rennan MF. Preoperative staging of esophageal cancer: comparison of endoscopic US + dynamic CT. Radiology 1991;181: Vilgrain V, Mompoint D, Palazzo L, et al. Staging of esophageal carcinoma: comparison of results with endoscopic sonography and CT. JR 1990;155: Rice TW, oyce G, Sivak MV. Esophageal ultrasound and the preoperative staging of carcinoma of the esophagus. J Thorac Cardiovasc Surg 1991;101: Wilke H, Preusser P, Fink U. Preoperative chemotherapy in locally advanced and nonresectable gastric cancer: a phase I1 study with etoposide, doxorubicin, and cisplatin. J Clin Oncol 1989;7131%26. DISCUSSION DR HRVEY I. PSS (ethesda, MD): When we are using computed tomograms for staging, size does not mean anything. Showing the consistency of a lymph node that is the same as the consistency of the esophageal cancer also really does not mean too much. How do you think endoesophageal ultrasound in the future is going to make an even bigger impact, possibly with other studies, to help with a histological diagnosis? re we content to stop here, or do we have to be more invasive on the preoperative assessment as we have been with lung cancer? DR RICE: Esophageal ultrasound is in its infancy. Currently we are familiarizing ourselves with the technique and the interpretation of the scans, and we are learning the power of esophageal ultrasound. What it has shown us to date is that the methods we have used in the past to assess the response to preoperative chemotherapy have been inadequate. We must realistically consider what we are subjecting our patients to. Preoperative chemotherapy and radiation therapy probably have not had an impact on survival of patients with esophageal carcinoma. We may learn this on completion of a phase I11 study, but at present, with the prediction of esophageal ultrasound, we can say that preoperative chemotherapy, at least the protocol used in the study, was totally ineffective. There is no need to wait until the survival data are complete. I think esophageal ultrasound will be crucial in the preoperative staging of patients with esophageal carcinoma. In the future, esophageal ultrasound may help direct therapy. I think it will be of utmost importance in the assessment of the effects of preoperative therapy.