Surgical Management of Lung Metastases: Selection Factors and Results
Surgical Management of Lung Metastases: Selection Factors and Results
Lung parenchyma is the most common site of metastatic disease
other than the liver for tumors that drain into the portal system
(Table 1). Of patients dying from metastatic disease to the lung,
20% to 25% may have no other evidence of metastases [1,2]. In
some instances, such as osteogenic and soft-tissue sarcomas in
adults, isolated pulmonary metastases represent the first recurrence
in 50% of patients [3,4].
Isolated case reports, followed by larger, more selected series,
have defined a set of criteria for patient suitability for surgical
resection of pulmonary metastases. Initially, patients with a
single metastasis and long tumor-free intervals were selected.
However, over time, the survival advantage of performing resections
in patients with unilateral multiple disease, bilateral multiple
disease, and even multiorgan disease, including the lung and liver,
have further extended the indications.
Frequently, chemotherapy and radiation therapy are used for palliation,
but only surgical extirpation, with or without multimodality therapy,
can realize a cure in selected instances. Combined studies confirm
an overall 5-year survival of 30% with low operative morbidity
In patients with a history of malignancy, the presence of new
pulmonary lesions should suggest the possibility of metastatic
disease. Once new nodule(s) are identified, a short, but important
list of criteria must be considered prior to resection. Even though
surgical metastatectomy has been performed for over 40 years,
the criteria have not changed significantly . Although there
are exceptions to every rule in medicine, the following standards
should be considered:
- The primary tumor site should be free of disease.
- Metastases should be confined to the lungs.
- The number of nodules should be "reasonable" (depending
on the histology).
- Chemotherapy is not curative.
- The patient can tolerate surgical therapy.
Standards, such as a specific doubling time, histology of the
primary tumor, and length of disease-free interval are variable.
However, these factors also should be considered in evaluating
resectability in the individual patient.
Attempts to rule out a lung primary carcinoma should be made,
as this would impact on the type and extent of operation recommended.
A history and physical examination, with evaluation of symptoms
and smoking status (and thus risk), are essential. Characteristics
of the nodule(s) on the chest radiograph and CT scan should suggest
either a primary or a secondary carcinoma. Most patients with
metastatic lung lesions are asymptomatic; however, this may also
be true for those with early-stage lung cancer. The presence of
hemoptysis, cough, or pain would more likely be associated with
a primary malignancy than a metastasis.
McCormack and Martini of Memorial Sloan-Kettering Cancer Center
found that a solitary pulmonary lesion was most likely a metastasis
if the primary was a sarcoma or melanoma, and a new primary if
the other malignancy was in the head and neck or breast; if the
previous lesion was of gastrointestinal or genitourinary origin,
however, a secondary or primary carcinoma was equally likely .
Although the majority of metastatic nodules are identified on
initial or follow-up chest x-rays, a chest CT scan is required
to obtain a more sensitive assessment of the number and characteristics
of the nodules. A chest x-ray may detect a 1-cm nodule, but a
CT scan may identify nodules 3 mm in size. With this increased
definition of nodules, specificity decreases, such that more benign
lesions, such as intraparenchymal lymph nodes and granulomas,
are revealed as possible metastatic abnormalities.
In a recent review of 144 patients undergoing resection of lung
metastases from colorectal carcinoma, the CT report differed from
pathologic findings in 30 (42%) of 72 patients. Among patients
with one or two radiologically imaged lesions, 12 patients had
fewer cancers (some lesions were benign) identified at surgery
and 18 patients had a higher number of malignant lesions; this
translated into an error rate of 28% . Despite these minor
limitations, CT scanning of the chest is the most sensitive noninvasive
test for preoperative assessment. In the majority of situations,
preoperative diagnosis is unnecessary, but it may be obtained,
if desired, by needle or transbronchial biopsy.
Other Imaging Modalities
Positron emission tomography (PET) is an emerging technology that
shows metastatic lesions as abnormal metabolizers of glucose,
compared to surrounding normal tissues. Current PET scans can
identify malignant nodules > 5 mm in size. Although this modality
has demonstrated excellent accuracy for diagnosing metastatic
melanoma, its sensitivity and specificity has not been confirmed,
particularly for other histologies . We anticipate that, in
the future, PET may provide a more definitive examination and
possibly supplant the CT scan.
Magnetic resonance imaging offers no advantage over CT for the
routine evaluation of metastatic pulmonary nodules.
The evaluation of a patient for resection of pulmonary metastases
is similar to standard prethoracotomy assessments. These procedures
are performed under general anesthesia, and therefore, the patient
must have adequate cardiopulmonary status. Resection of pulmonary
metastases usually does not necessitate a lobectomy and even less
frequently requires a pneumonectomy; therefore, more borderline
pulmonary function results may be acceptable. Clearly, the location
and number of lesions is critical in planning the type of resection
required. Postoperatively, the patient should have no less than
a FEV1 (forced expiratory volume in 1 second) of 800 to 1000 cc;
with a preoperative DCO (diffusing capacity) > 50%; MMV (a
maximum voluntary ventilation) > 50%; and a PCO2 < 45.
Control of the primary site and the absence of further metastatic
disease are confirmed by blood chemistry, brain MRI, bone scan,
and CT or MRI of the primary site. For the majority of metastases,
confirmatory tissue diagnosis of the lesion is unnecessary prior
Type of Incision
The type of incision depends on the location and extent of the
lesions. If the lesions are peripheral and bilateral, a median
sternotomy may be appropriate. This has the advantages of less
postoperative pain, more rapid early recovery of pulmonary function,
and better identification of bilaterality that may not have been
detected in the preoperative evaluation . Of patients demonstrating
unilateral metastases on CT scan, 45% have bilateral disease at
A posterior lateral thoracotomy, particularly on the left, may
be preferable for a lower lobectomy or posterior wedge resection
and for adequate resection of the diaphragm that is contiguous
with the metastasis. In these instances, with information known
from the preoperative CT scan concerning bilaterality and diaphragmatic
extension, staged thoracotomies would be more appropriate. Also,
if a previous median sternotomy has been performed, repeated access
may be more difficult, and thoracotomies may be preferred for
further resections. The transverse or clamshell incision should
also be considered for patients who have bilateral disease, especially
in the lower lobes [11,12].
Video-Assisted Thoracoscopy Surgery
Video-assisted thoracoscopy surgery (VATS) is a newer surgical
approach that may have been overutilized recently. Most surgeons
are concerned that the video-assisted technique does not allow
for manual palpation of the lungs in order to identify small (<
5 mm) lesions seen on CT scan. As described above, a retrospective
review demonstrated a 28% error rate of CT scans, compared with
pathologic results . This suggests that VATS might have missed
these other lesions that were identified at the time of open thoracotomy.
The results of an ongoing randomized trial examining the efficacy
of VATS will hopefully settle this issue.
The claim that patients undergoing VATS-wedge resections have
reduced morbidity, cost, and hospital stay has yet to be verified
with certainty.13 Moreover, the incidence of local recurrences
at the resection margins, in the pleural space, and in port sites
may be higher after thoracoscopic resection of metastatic disease.
It is extremely important not to spill tumor contents into the
pleural space. This can occur when the stapling device is used,
and may happen without the surgeon's knowledge because of a contaminated
resection margin. Once a pleural spill occurs, patients may progress
to pleural metastatic disease, which usually is not curatively
At present, we recommend an open resection, preferably with simultaneous
bilateral dissections, in clinically qualified patients. Video-assisted
thoracoscopy resection is not advocated as conventional therapy,
but may be reserved for the occasional patient who has limited
pulmonary function and a single metastasis with a long disease-free
The goal of resection is to remove all identified or palpated
disease with the minimal volume of normal surrounding parenchyma.
This is usually achieved by a wedge resection, Perelman resection
(precision excision with either cautery or laser), lobectomy,
or pneumonectomy, depending on the location of the nodule. An
adequate resection margin of 0.5 cm is preferred, but unfortunately,
is not always feasible. A cytologic touch preparation of the resected
margin should always be performed if the margin is suspicious
When the primary and metastatic lesions present simultaneously,
surgical treatment of the primary should be done initially. If
a curative resection of the primary can be performed successfully,
metastatectomy should be carried out approximately 4 weeks postoperatively.
If resection of the primary is noncurative, usually there is no
need to resect the pulmonary metastases.
Recently, extensive metastatic disease isolated to the lungs has
been perfused with chemotherapeutic agents in animal and clinical
models [14,15]. The techniques no longer require systemic circulatory
assist and can be performed bilaterally through a mid-line sternotomy
(HI Pass, MD, personal communication, October 1995). Although
presently experimental, the rationality and efficacy of organ
perfusion based on the liver perfusion model warrants further