Mediastinoscopy. A review of anatomical relationships and complications

Successful management of severe pulmonary artery injury during mediastinoscopy

Although many reports have shown that mediastinoscopy is an extremely safe procedure, the most serious complication that can occur is hemorrhage from the large vessels in the mediastinum. A patient with severe pulmonary artery injury during mediastinoscopy that was successfully repaired through a right posterolateral thoracotomy is reported. Digital compression based on the anatomy was highly effective in controlling the hemorrhage caused by a longitudinal tear.

Management of major hemorrhage during mediastinoscopy

OBJECTIVE

The management of major hemorrhage complicating mediastinoscopy is not well described. We reviewed our experience to determine the frequency, optimal management strategy, and outcome of these injuries.

METHODS

A retrospective review of all mediastinoscopies performed at a single institution during a 12-year period (January 1990-January 2002) was performed. Major hemorrhage was defined as that requiring exploration for definitive control.

RESULTS

During the study period, 3391 mediastinoscopies were performed. Fourteen patients (0.4%) experienced major hemorrhage. Most patients (12/14) had non-small cell lung cancer, and only 1 patient each underwent preoperative radiation, repeat mediastinoscopy, or extended mediastinoscopy. The most common biopsy site (4/14 cases) resulting in major hemorrhage was the lower right paratracheal region (level 4R), and the most frequently injured vessels were the azygos vein and the innominate and pulmonary arteries. Initial control of hemorrhage was achieved through packing in 93% (13/14), and the most common initial approach for exploration was sternotomy (8/14). Four patients underwent a planned pulmonary resection after definitive control of bleeding. The median amount of blood transfused was 2 units (range 0-18 units). Postoperative complications occurred in 2 of 14 patients (14%). There were no intraoperative deaths, but 1 patient died postoperatively (1/14, 7% mortality). The median postoperative length of stay was 6 days (range 1-19 days).

CONCLUSIONS

Major hemorrhage during mediastinoscopy is an uncommon but potentially morbid event. Initial control can usually be achieved through packing. Subsequent management presents a technical challenge but can result in minimal morbidity and mortality.

Mediastinoscopy and mediastinotomy after cardiac surgery: are safety and efficacy affected by prior sternotomy?

BACKGROUND

The diagnostic and staging value of cervical mediastinoscopy is well established. Left anterior mediastinotomy is of further value in assessing left upper lobe tumors. However the efficacy and safety of both these procedures after median sternotomy for cardiac surgery is unknown.

METHODS

We undertook a retrospective review of our experience of mediastinal exploration by cervical mediastinoscopy with or without left anterior mediastinotomy in patients with prior sternotomy between 1980 and 2001.

RESULTS

During this period 28 patients (25 male and 3 female; mean age, 63 +/- 10 years), all with prior sternotomy for cardiac surgery (14 had left internal mammary artery graft), underwent mediastinal exploration. The mean interval between sternotomy and mediastinal exploration was 7.2 +/- 5.1 years. Additionally, 3 patients also had superior vena cava obstruction. Cervical mediastinoscopy was performed in all 28 patients and additionally left anterior mediastinotomy was undertaken in 7 of 28 patients (4 with left internal mammary artery graft). Indications for exploration were staging of lung cancer in 22 patients (cervical mediastinoscopy, n = 22; left anterior mediastinotomy, n = 7) and diagnostic biopsy of mediastinal mass in 6 patients (cervical mediastinoscopy, n = 6). Thorough mediastinal assessment was possible in all 28 patients. In the 22 patients with lung cancer the median number of lymph node stations sampled during mediastinoscopy was 3 (range, 1 to 5). A specific diagnosis was obtained in 16 patients (metastatic lung cancer, n = 10; lymphoma, n = 3; sarcoidosis, sinus histiocytosis, and metastatic melanoma, n = 1 each). The other 12 patients with negative findings underwent pulmonary resection and only 1 of 12 (8%) patients had unexpected N2 disease, a similar proportion to our overall experience with lung cancer. There were no operative complications.

CONCLUSIONS

Prior sternotomy for cardiac surgery does not compromise the efficacy and the safety of mediastinoscopy and mediastinotomy.

Bronchoscopy and surgical staging procedures and their correlation with imaging

Bronchoscopy, computed tomography (CT) and surgical staging procedures are complimentary methods of investigating patients with lung cancer. CT has been shown to be of value prior to bronchoscopy in the investigation of haemoptysis and malignancy, with excellent correlation between the detection of disease within the large airways on CT and direct visualisation at bronchoscopy. The utility of CT has been further increased by the development of multislice scanners with the generation of volumetric data enabling multiplanar image acquisition. Additionally the advent of CT co-registered with positron emission tomography will play an important role in guiding the choice of surgical staging procedures The increasing use of multidisciplinary medical care requires radiologists to have a greater understanding of the abilities and limitations of both bronchoscopy and surgical staging procedures in evaluating disease demonstrated on imaging.

99mTC-tetrofosmin scintigraphy in lung carcinoma staging and follow-up evaluations

BACKGROUND

99mTC-tetrofosmin recently has emerged as a new radiopharmaceutical for cancer visualization. In this study, the authors have investigated, for the first time in a comprehensive way, its ability to assess lung carcinoma dissemination and progression.

METHODS

A 99mTC-tetrofosmin scan was incorporated into the pretreatment and posttreatment diagnostic workup of all lung carcinoma patients seen in a second referral institution for a province of 500,000 inhabitants during the years 1998 and 1999. Sixty-one patients, strongly suspected of lung carcinoma were photon-scanned; 21 of them were rescanned after completion of their front-line treatment. Eleven patients eventually underwent surgery, and 3 others underwent mediastinoscopy. Both planar and single photoemission computed tomography thoracic views were obtained. Images for the whole body also were acquired.

RESULTS

All 57 patients whose lung carcinoma was pathologically confirmed showed accumulation of the radiotracer (100% sensitivity). However, three of the four nonmalignant lesions were also 99mTC-tetrofosmin positive. 99mTC-tetrofosmin scan was highly sensitive for the detection of the T0-T2 disease (97% sensitivity) and highly specific for the N0-N1 disease (83% specificity). In the 16 pathologically staged mediastina, sensitivity, specificity, and accuracy rates were 73%, 100%, and 81%, respectively. 99mTC-tetrofosmin scan correctly detected most skeleton (9 of 10) and brain (5 of 7) metastases. The treatment response evaluation made with 99mTC-tetrofosmin corresponded to the clinical estimate in almost half of the sample.

CONCLUSIONS

This study shows that 99mTC-tetrofosmin scan is a relatively accurate method for lung carcinoma evaluation. The authors' preliminary data exclude, however, that noninvasive diagnostic efficiency might be substantially increased by a scintigraphy with 99mTC-tetrofosmin. More studies are needed for a better understanding of the real value of this technique.

Surgical management of early stage lung cancer

For patients with lung cancer, the greatest hope for cure rests with patients with early stage disease. Surgery has been the standard of care for this group with the best 5-year survival of only 65% being achieved in patients with earliest pathologic Stage IA disease. Using strategies gained from the management of patients with advanced disease, clinicians are investigating the use of perioperative chemotherapy and radiotherapy to improve survival. In addition, biologic and molecular markers are being evaluated to assist in predicting prognosis and to identify those patients at increased risk for recurrent disease. Postoperative surveillance of patients using helical computed tomography (CT) scanning is being investigated to detect early recurrences and second primary lesions. With such treatment and management plans on the horizon, the prognosis of patients with early stage non-small cell lung cancer (NSCLC) may be improved.

[Descriptive study of complications caused by mediastinoscopy]

To analyze the surgical and non surgical complications of mediastinoscopy in a series of 200 consecutive patients. Retrospective study of all surgical patients between 1 January 1994 and 1 May 1997. Any complication presenting between time of surgery and patient release is analyzed. Complications were seen in 8 out of 200 cases (4%). In three cases, there were lesions in neighboring structures (recurrent nerve, bronchial artery and innominate artery). The rest were non surgical complications (3 cases of arrhythmia, 2 of prolonged mechanical ventilation). One male patient (0.5%) died from cerebral infarction, probably as a result of arterial occlusion needed to suture damage to the innominate artery. Superior win cava syndrome affected 20% (1 in 5) and morbidity was 60% (3 in 5). Morbidity involving both medical and surgical complications in this series is higher than that reported elsewhere in the literature, in series for which non surgical complications go unreported.

Anti-CEA immunoscintigraphy and computed tomographic scanning in the preoperative evaluation of mediastinal lymph nodes in lung cancer

BACKGROUND

Thoracic computed tomography (CT) provides most of the staging information needed before operation for lung cancer and can reduce the number of exploratory thoracotomies. In recent years a new immunoscintigraphic technique with anti-carcinoembryonic antigen (CEA) monoclonal antibodies has been shown to be effective in lung cancer staging. This study compares the yields of CT scans and immunoscintigraphy in the preoperative evaluation of the medistinal lymph nodes of patients with non-small cell lung cancer.

METHODS

One hundred and thirty one patients believed on clinical grounds to have a operable non-small cell lung cancer were photoscanned with the indium-111 labelled F(ab')2 fragments of the antibody FO23C5. Both planar and single photoemission computed tomography (SPECT) thoracic views were recorded. CT scan of the thorax, abdomen, and brain were obtained in all patients. Seventy of the patients eventually underwent surgery, an additional seven underwent mediastinoscopy or mediastinotomy, and a further 10 had both cervical exploration and thoracotomy. Pathological evaluation of the mediastinal nodes was available in all 87 patients, but in only 80 of them was the diagnosis of lung cancer eventually confirmed.

RESULTS

The diagnostic accuracy of planar immunoscintigraphy, SPECT immunoscintigraphy, and CT scanning for N2 disease was 76%, 74%, and 71%, respectively. The corresponding sensitivity and specificity rates were 45%, 77%, 64% and 88%, 72%, and 74%. These were not significantly different.

CONCLUSIONS

This study shows that anti-CEA immunoscintigraphy has no advantage over conventional CT scanning in assessing mediastinal lymphoadenopathy in patients with lung cancer. CT scanning remains the gold standard test in these patients.

Cerebrovascular accident complicating extended cervical mediastinoscopy

The safety and efficacy of staging cervical mediastinoscopy is well established. Extended cervical mediastinoscopy has been proposed as a safe and effective method of staging left upper lobe lung cancers. We report a case of cerebrovascular accident complicating extended cervical mediastinoscopy.

Anti-CEA immunoscintigraphy might be more useful than computed tomography in the preoperative thoracic evaluation of lung cancer. A comparison between planar immunoscintigraphy, single photon emission computed tomography (SPECT), and computed tomography

While a clinical, plain radiographic, and bronchoscopic assessment yields most of the essential information needed in lung cancer, computed tomography (CT) of the thorax provides diagnostic information previously unobtainable, potentially capable of reducing the number of explorative thoracotomies. In a few recent studies, immunoscintigraphy with anti-carcinoembryonic antigen (anti-CEA) monoclonal antibodies (MA) has shown remarkable staging potential. To compare the diagnostic accuracy of the two techniques, we photoscanned with indium-111 (111In)-labeled-F(ab')2 fragments of the murine anti-CEA MA FO23C5 45 patients, who were pathologically assessed for possible loco-regional extension of lung cancer. Both planar and single photo emission computed tomography (SPECT) images were obtained. Additionally, CT of the thorax (contiguous CT slices, 10 mm thick, from the lung apices to the upper abdomen), and other routine tests of preoperative evaluation were obtained. On the basis of 37 (N1, T3, and T4), 38 (N2), and 12 (N3) pathologically documented sites, an accuracy of 65, 76, 92, 78, and 86 percent (SPECT images), and 62, 68, 42, 78, and 84 percent (CT images) was calculated (figures are relevant to N1, N2, N3, T3, and T4 disease, respectively). Thus, both techniques shared a significant margin of error in almost all the categories of evaluation; however, immunoscintigraphy showed equivalent, and, in the lymph node assessment, superior results to CT. A marginal improvement of diagnostic accuracy was recorded combining the three techniques in one case (SPECT plus planar immunoscintigraphic images), while there was no benefit in any possible integration of CT and immunoscintigraphic images. In patients with peripheral nonsquamous cell cancers, the accuracy of anti-CEA immunoscintigraphy was of 90 percent or higher. Variations in the modality of performing immunoscintigraphy, such as changes in the dose of antibody fragments to be injected, in the percentage of radiolabeling, or in the time of imaging, affected the quality of immunoscintigraphic series, and the consequent interpretation of findings. At the present time, there are very few reliable tests capable of selecting patients to proceed directly to thoracotomy or to receive some intermediate surgical test, such as a prior mediastinoscopy. Traditionally, CT has been this type of "filter-test." If current findings will be confirmed in future studies, anti-CEA immunoscintigraphy might replace CT in the evaluation of particular subgroups of patients, such as patients with peripheral nonsquamous cell bronchogenic carcinoma.

Imaging lung cancer by scintigraphy with Indium 111-labeled F(ab')2 fragments of the anticarcinoembryonic antigen monoclonal antibody FO23C5

BACKGROUND

Anticarcinoembryonic (CEA) monoclonal antibodies are able to react specifically with the antigen and have the potential for the detection of CEA-bearing tumors.

METHODS

The authors photoscanned with indium 111 (111In)-labeled F(ab')2 fragments of the murine CEA monoclonal antibody FO23C5 63 patients with a newly diagnosed and pathologically documented bronchogenic carcinoma. Planar dual views of the thorax, abdomen, and brain were acquired between the 24th and 144th hour after the radiotracer injection. Patients had a complete pretreatment workup, which included a routine multiorgan computed tomography (CT) scan, and the determination of the serum and tissue CEA concentration. All patients were followed up clinically and radiologically. Nineteen needle aspirations and biopsies, 23 surgical explorations, and 4 mediastinoscopic studies yielded 121 pathologically documented sites of reference.

RESULTS

Fifty-seven of 63 scans were positive for the primary tumor (sensitivity, 0.90). The uptake of the radiotracer correlated significantly with the intensity of tissue CEA expression (Spearman R [Rs], 0.25; P less than 0.05), but not with the serum CEA level or with the histotype. Overall, the sensitivity of the anti-CEA immunoscintigraphy (IS) for the N1, N2, N3, T3, T4, and M1 disease (1987 International Union Against Cancer [UICC] staging classification) was 0.67, 0.64, 0.62, 0.31, 0.29, and 0.86, respectively. Corresponding values of specificity were 0.67, 0.81, 0.90, 1, 1, and 0.93; accuracy values were 0.67, 0.71, 0.85, 0.71, 0.76, and 0.92. The authors limited the analysis to all of the pathologically documented sites and obtained slightly superior values but no meaningful differences. The stage derived from IS readings was correct in 33 patients. The same figure was obtained after an initial clinical workup, which included physical examination, laboratory routine tests, chest radiographs, bronchoscopy, and any diagnostic procedure indicated by those tests.

CONCLUSIONS

Anti-CEA FO23C5-F(ab')2 fragments are not yet "magic bullets" for perfect diagnoses; however, their staging potential seems to be remarkable. Technical improvements, single-photo emission CT, and the use of such fragments in combination with other imaging techniques might enable researchers to further improve the current results.

Complications of mediastinoscopy

From 1968 to 1987, mediastinoscopy was performed on 2021 consecutive patients. The procedure was carried out in 35.7 per cent on patients with a pulmonary infiltrate and in 28.8 per cent with bronchogenic carcinoma. Mediastinoscopy was positive in 54 per cent of the cases. The total number of complications was 2.3 per cent (47/2021). There were four cases with major haemorrhage, three tracheal ruptures and three wound infections. Recurrent nerve paralysis was verified in three patients and pericardial rupture in two patients. The total of major complications was 0.5 per cent (10/2021). There was no deaths in this series. In the authors' opinion, mediastinoscopy can be regarded as an effective and safe procedure in the hands of an experienced surgeon.

Mediastinal assessment for staging and treatment of carcinoma of the lung

Invasive diagnostic procedures for mediastinal assessment, such as mediastinoscopy, are necessitated by the importance of staging lung cancers, both to plan the treatment and to estimate the prognosis. Other noninvasive techniques may complement or be substituted for mediastinoscopy under certain specific clinical settings. Thus with the introduction of newer diagnostic technologies, such as computed axial tomography, the strategy for mediastinal assessment should be continually reevaluated. In this review, the diagnostic sensitivity, specificity, and overall accuracy of various techniques reported in the literature are examined to elucidate their current roles in assessing the mediastinal involvement in patients with lung cancer.

Role of computerized tomography in assessment of the mediastinum in patients with lung carcinoma

Over a 28 months period, 75 patients with pathologically confirmed lung carcinoma had computerized tomographic scanning of the mediastinum as a staging procedure. The scan had a sensitivity of 85 percent and a specificity of 89 percent for mediastinal lymph node metastases, making routine mediastinoscopy unnecessary. Tissue confirmation of most positive scans is advisable since false-positive results can deny some patients potentially curative resection. Invasive staging should also be employed in the face of negative scans in selected patients. Computerized tomographic scanning provides anatomic information that is helpful in directing these invasive staging procedures.

Computed tomography for evaluation of mediastinal lymph nodes in lung cancer: correlation with surgical staging

Computed tomography (CT) of the chest (late model) was done preoperatively in 56 candidates for resection of lung cancer. Precise borders for each node region were defined by the American Thoracic Society modification of the classification of the American Joint Committee for Cancer Staging and were used to "map" nodes seen on CT and nodes removed surgically. Metastatic involvement of mediastinal nodes was proven by mediastinoscopy in 11 patients; nodes were removed from multiple regions at thoracotomy in 45 patients. The mediastinum was clearly delineated by CT in 46 patients with determinate scans and was judged normal in 32 (CT-negative scans) and abnormal in 14 (CT-positive scans). A node was considered metastatically involved if it measured greater than 1.5 cm in diameter. Positive nodes were found at surgical staging in 3 of 32 patients with CT-negative scans and in all patients with CT-positive scans. Thus, for the 46 patients with determinate scans, sensitivity was 82%, specificity was 100%, and accuracy (true positive and true negative) was 93%. The high accuracy of CT in these patients suggests that mediastinoscopy is not necessary before thoracotomy in the patient with a CT-negative scan, but that for the patient with a CT-positive or CT-indeterminate scan, the indications for mediastinoscopy remain the same.

Evaluation of tomography and mediastinoscopy for the detection of mediastinal lymph node metastases

In a prospective study of 88 patients seen consecutively with proven or suspected bronchial carcinoma, the validity of x-ray tomography and routine mediastinoscopy was tested for the detection and evaluation of mediastinal lymph node metastases. Positive mediastinum was defined as malignant tissue found in the mediastinum and negative mediastinum as mediastinoscopy with negative results plus a negative intraoperative mediastinal lymph node dissection. Thirty-four patients were eliminated from the analysis because carcinoma was not found or because mediastinal evaluation was incomplete by these criteria. Twenty-eight of the remaining 54 patients had mediastinal metastases. Sensitivity was 67% for tomography and 79% for mediastinoscopy. Specificity was 92% for tomography and 100% for mediastinoscopy. The differences were not significant. Sixty-six of 85 mediastinoscopies were unnecessary or unhelpful in the decision to exclude a patient from surgical intervention. Among 19 patients with lesions presumed to be inoperable based on results of mediastinoscopy (i.e., perinodal metastatic growth suspected by palpation or histologically proven), 14 patients had positive tomographic scans and 1 could not be evaluated radiographically because of right upper lobe atelectasis. We conclude that tomography of the upper mediastinum should be used to select patients for mediastinoscopy.

Preoperative staging of lung cancer: accuracy of computed tomography versus mediastinoscopy

Forty-four patients coming to surgery for carcinoma of the bronchus underwent preoperative staging of the mediastinum by computed tomography (CT scanning) and surgical exploration of the mediastinum by cervical mediastinoscopy or left anterior mediastinotomy or both. Where mediastinal nodes were affected the sensitivity and specificity of computed tomography was inferior to that of mediastinoscopy (57% and 85% versus 71% and 100%). The sensitivity of computed tomography in predicting mediastinal invasion was superior to that of mediastinoscopy (77% v 46%), especially in the case of lower-lobe tumours (67% v 17%). Mediastinoscopy had the considerable advantage of 100% specificity. In the assessment of hilar lymphadenopathy computed tomography had a sensitivity of 38% and a specificity of 64%. In cases where computed tomography showed a normal mediastinum or enlargement of the hilar glands only, mediastinal exploration conferred no additional information and could have been omitted. A computed tomography scan showing mediastinal abnormality is an indication for mediastinoscopy and not a contraindication to surgery. In 23 patients computed tomography showed some abnormality of the mediastinum, confirmed at mediastinoscopy in 12 cases. The remaining 11 patients underwent thoracotomy, resection being carried out in nine. Postsurgical staging showed that six of these tumours were N0 lesions without invasion; in two further N0 cases there was a minor degree of mediastinal invasion which did not prevent resection, and the remaining tumour was N1 without invasion.

The value of radiographic and computed tomography in the staging of lung carcinoma

A prospective double-blind study was undertaken to compare computed tomography (CT) and conventional radiographic tomography (RT) in the staging of lung carcinoma. Seventy-five patients had CT and RT of the mediastinum and hilum prior to operation. The presence or absence of metastasis to lymph nodes documented at the time of operation was the standard applied to the studies. CT correctly predicted the presence or absence of mediastinal lymphadenopathy in most cases (sensitivity 91%, specificity 94%), while RT was less helpful (sensitivity 61%, specificity 86%). Metastatic mediastinal lymph nodes in those patients with false negative CT and RT studies averaged only 0.8 cm in diameter, probably accounting for the negative radiographic findings. Both CT and RT had poor predictive values in detecting hilar lymphadenopathy (sensitivity 73% and 47%, specificity 87% and 72%, respectively). The predictive value of CT in the evaluation of mediastinal lymphadenopathy equaled that of mediastinoscopy or mediastinotomy. When CT of the mediastinum demonstrates no lymphadenopathy, invasive staging can be deferred for definitive thoracotomy. Since false positive values were seen with both CT and RT scans of the mediastinum (4% and 8%, respectively), invasive staging will still be necessary in those patients with positive studies.

Noninvasive evaluation of mediastinal metastases in bronchogenic carcinoma: a prospective comparison of chest radiography and gallium-67 scanning

Evaluation of regional node involvement in patients with bronchogenic carcinoma is a crucial step in determining therapy and prognosis. Mediastinoscopy has been recommended for staging all potentially operable cases, but technical limitations and the need for anesthesia make this impractical. Gallium-67 scanning and chest radiography were prospectively compared as noninvasive means of evaluating spread to regional nodes in 75 patients with bronchogenic carcinoma in whom histologic evaluation of hilar and mediastinal nodes was performed. Gallium scanning was more accurate than chest radiography in assessing regional nodes (overall accuracy 85.3% vs. 56%, P less than 0.05). When positive, both procedures correctly indicate malignant involvement of regional nodes (85% vs. 87.3%). A negative gallium scan, however, was significantly more accurate in predicting the absence of such involvement (80% vs. 40%, P less than 0.01). Gallium scanning appears to be a reliable, noninvasive means of assessing mediastinal spread of bronchogenic carcinoma and when used in conjunction with radiographic findings, allows selection of appropriate patients for surgical staging procedures.

The importance of mediastinal lymph node invasion by pulmonary carcinoma in selection of patients for resection

Resection results from 417 consecutive patients operated on between January 1, 1964, and December 30, 1969, were analyzed in March, 1976. This period was chosen to allow a five-year follow-up. The results of resection in 56 patients with invaded mediastinal nodes are reported. Mediastinoscopy to assess resectability was not used for any of the 417 patients. Our low incidence of mediastinal node invasion (56 out of 417, or 13.4%), a resectability rate of 97.4%, and a hospital mortality of 2.8% for resection of advanced carcinoma suggest that routine mediastinoscopy prior to resection is not necessary. Traditional methods of preoperative assessment and the use, when indicated, of extended resection for patients with mediastinal node invasion result in worthwhile salvage of patients with invaded mediastinal nodes.

Is mediastinoscopy necessary in the evaluation of lung cancer?

The records of 96 consecutive patients who underwent mediastinoscopy and were ultimately shown to have bronchogenic carcinoma were reviewed. Indirect tests for mediastinal tumor metastases in these patients included bronchoscopy and chest roentgenograms in all 96, mediastinal laminagrams in 65, esophagograms in 27, carinal biopsy in 23, bronchograms in 5, pulmonary angiograms in 5, azygograms in 2, and aortograms in 2 patients. Of the 43 patients in this series in whom all indirect tests revealed no metastases, mediastinoscopy showed nodal involvement in 11 (28%), who were thus spared unnecessary thoracotomy. On the other hand, if negative mediastinoscopy had not cast doubt on the validity of indirect tests that seemed to show metastases, an operation might actuallly have been denied to 14 patients who were ultimately proved to have anatomically resectable disease.