Introduction

Pulmonary embolism (PE) has traditionally been considered a discrete syndrome that requires its own specific methods of diagnosis¹. Although a number of tests have been used for diagnosis of PE, the ventilation/perfusion lungs scans id the most frequent initial imaging study requested by clinicians as a noninvasive diagnostic procedure for evaluation of PE ²,³,⁴,⁵,⁶.

The association between malignancy and thrombosis has been recognized since Trousseau described the syndrome o0f recurrent migratory thrombophlebitis in patients with cancer over a century ago ⁷. Since that time, the presence of a malignant neoplasm was considered to be an important predisposing factor for PE ⁸,⁹,¹⁰,¹¹,¹²,¹³,¹⁴,¹⁵.

The diagnosis of PE in patients with malignant disorders is difficult because the presentation of PE may mimic other complications of malignancy such as metastasis, pneumonia, pleural effusion, etc. ; and due to the inherent anxiety PE may tend to be overdiagnosed.

The diagnosis of PE in cancer patients requires high degree of accuracy to avoid the potentially life-threatening complications of long-term treatment with anticoagulants⁵,¹⁶.

Clinical and radiological evaluation of patients may suggest the possibility of pulmonary embolism, but cannot reliably make or exclude the diagnosis. The V/Q scan reflects the characteristic physiologic derangement in pulmonary thromboembolism (lack of perfusion of ventilated lung segments), it is safe and more accurate than any other noninvasive method¹⁷.

Pulmonary angiography remains the most specific and sensitive diagnostic procedure available to diagnose pulmonary embolism. However, it carries significant morbidity and mortality ¹⁸.

The purpose of this study was to review the incidence, clinical characteristics, and radiological findings of suspected PE in malignant disorders as well as the accuracy of V/Q lung scan as a diagnostic tool.

Results

We reviewed 293 cancer patients who had V/Q lung scan done at M.D. Anderson Cancer Center during a 18 month period, due to suspicion of pulmonary embolism. Of these 293, 158 were female (53.9%) and 135 male (46.1%), their ages ranged from 13 to 89 years with mean values 54.86 6 14.55 years.

As shown in Table I, only 29 out the 293 (9.9%) cases had high probability V/Q scan and 251 (85.6%) had either normal of low probability V/Q scan.

Figure 1

Table 1: V/Q lung scan results in 293 cancer patients with clinically suspected pulmonary embolism.

Table II reveals the V/Q lung scan results according to the different types of malignancy, Lympoma, breast CA, and pulmonary CA were the most frequent types of malignancies , while brain tumors (CNS CA) and melanoma (5cases) & multiple myeloma (2 cases) were the least frequent cancers. The ratio high probability scans/total scans (incidence of PE per type of malignancy) was maintained in all the different categories.

No significant relationship was found between the presence of metastasis and the V/Q scan probability for PE (table III).

Figure 2

Table 2: V/Q lung scan result and its relationship with type of malignancy.

Figure 3

Table 3: V/Q lung scan results in suspected PE in relation to detection of metastasis.

No significant relationship was found between the presence of metastasis and the V/Q scan probability for PE (Table III).

The clinical characteristics (symptoms and signs) and their association with high probability V/Q scans are shown in Table IV. Dyspnea and chest pain were the symptoms that revealed highly significant association with PE.

Among the signs, tachycardia, crackles, tachypnea, hypotension, and the presence of deep venous thrombosis (DVT) were highly significant related to high probability V/Q scan.

Figure 4

Table 4 : Frequencies of clinical characteristics of malignant patients with suspicion of PE in relation to V/Q lung scan probabilities for PE.

Figure 5

Table 5: Chest roentgenographic findings in our population, in relation to V/Q scan results.

The different chest roentgenographic finding are llustrated in Table V. Eighty two cases (27.9%) had normal CXR and five of them had high probability for PE, this result was highly significant (p< 0.001). Atelectasis either right or left and segmental or larger was the most common radiological abnormality which was also significant (p<0.05). Pleural effusion either right or left was the least frequent CXR finding and the rest of the findings were non significant. Other radiological findings did not show relationship with the V/Q probability.

When the reasons for requesting a V/Q lung scan were complied and compared with results, dyspnea, chest pain, hypoxemia, and DVT were significantly associated with the highest diagnostic yield, as depicted in Table VI.

Figure 6

Table 6 : Reasons for requesting a V/Q scan and results.

Figure 7

Table 7: Relation of V/Q lung scan prob. For PE and therapy that patients Received within the last three months.

As illustrated in Table VII, chemotherapy and surgery disclosed a highly significant correlation with the V/Q lung scan probability (p<0.001).

Pulmonary angiography was done in five patients (one with high probability V/Q lung scan and four with intermediate probability) and it was positive in three (one high and two with intermediate probability).

Autopsy was performed in six patients (3 normal and 3 with low probability V/Q lung scan), and one, who had low probability V/Q, revealed autopsy findings compatible with PE (which had low V/Q).