Advances in Peptide-Receptor Imaging of Thromboembolic Disease and Lung Cancer

At an industry-sponsored symposium held in conjunction with CHEST 2002, the annual meeting of the American College of Chest Physicians, six leaders in physics, pulmonary medicine, and nuclear medicine presented the latest information on peptide-receptor imaging and its uses in patients with confirmed or suspected thromboembolic disease or lung cancers. Topics included the evolution of the diagnostic approaches to deep venous thrombosis (DVT) and pulmonary embolisms (PE), use of the Tc99m apcitide in diagnosis of acute blood clots, the early diagnosis of lung cancer and the TC 99m depreotide to characterize the solitary pulmonary nodule.

This program was supported by an unrestricted educational grant from Berlex Laboratories.

Diagnostic Approach to Deep Venous Thrombosis/Pulmonary

“Up to the mid-1980s, veno-grams were done frequently for the diagnosis of DVT,” said Paul D. Stein, MD, FCCP, of St. Joseph Mercy Oakland Hospital in Pontiac, MI. “Even though it is still considered the gold standard for diagnosis, their use has decreased to virtually zero.”

Venograms have largely been replaced by compression ultrasound (US) in the thigh. “Venous US is good in the thigh, not so good in the calf, good in symptomatic patients but less useful in those without symptoms,” said Dr. Stein. “Serial US is more effective, although insurance won’t pay for additional
studies.”

Other modalities are available for the diagnosis of DVT, but all have problems. Spiral CT (S-CT) following injection of contrast material into the veins of the lower extremities is highly sensitive but the contrast material causes discomfort and possible venous irritation. Magnetic resonance imaging (MRI) is sensitive and specific, but the enclosed environment and expense are limiting factors.

The gold standard for PE remains the pulmonary angiogram. There is some hesitation to order it related to its well-established 1% incidence of serious morbidity. V/Q scans yield a definitive diagnosis less than half the time.

“Serial US of the lower extremities of patients with suspected PE can serve as a surrogate for the diagnosis of PE, however lots of time we can’t obtain serial venograms,” said Dr. Stein. “Studies have shown that a negative D-dimer test and a low clinical probability can exclude PE in 99% of patients. Low clinical probability in combination with a non-diagnostic V/Q scan and a negative US also excludes PE.”

Contrast-enhanced S-CT is becoming more widely used. A European study noted sensitivity varied according to who read the films. Local readers, based on pulmonary angiography, found it sensitive in 88% of the patients as compared with 60% in the expert central-reader’s response to the same films (Herold CJ, et al. Radiology [abstract] 1998;209P: 299). The ANTELOPE study noted 69% sensitivity and a report from Mayo and others showed 87% sensitivity (van Strijen MJ, et al. Radiology (abstract) 1999;213P:127;Mayo JR, et al. Radiology 1997;205:447).

“There are wide variations from study to study,” said Dr. Stein. “We just don’t yet know the true specificity and sensitivity of S-CT for PE.”

Tc99m Apcitide in the Diagnosis of Acute Clot

Most diagnostic studies available are anatomic or morphologic. They do not tell anything about what happens to the clot.

“Because of their invasive nature, angiography and venography are being used less and technical expertise is declining,” said Lisa Moores, MD, FCCP, from the Walter Reed Army Medical Center in Washington, DC. “I’m not sure I have much faith in the read on an angiogram these days just because our radiologists aren’t doing them.”

Non-invasive studies rely on finding a nearly occlusive clot, making them less sensitive in asymptomatic patients. They do not differentiate between a
recurrent clot and an acute clot.

There is interest in biological methods to detect clots. Blood flow and pooling techniques have shown modest success. Radio-labeling fibrinogen has not been widely accepted due to problems separating background data from clot data.

This leads to interest in studying smaller peptides, such as technetium 99m (Tc99m) apcitide. Only activated platelets express the glycoprotein IIB/IIIA receptor. The Tc99m apcitide competes with fibrinogen for activated platelets giving a very nice signal-to-noise ratio. Acquisition only takes 10 minutes for the early phase, followed by a delayed phase 60 to 120 minutes later. Continued clinical use of the product now recommends just one set of images acquired at 120 minutes post-injection.

“We look for abnormal areas showing active binding of the radiolabeled product that persists over time,” said Dr. Moores.

There are a number of instances when Tc99m apcitide labeling could prove useful. For example, Dr. Moores discussed the case of a 64-year-old male with high fever, chills and abnormal chest x-ray not resolved with empiric treatment for pneumonia. The patient had had prior PE requiring a differentiation between a potential new/recurrent clot and an older residual perfusion defect.

“Tc99m apcitide showed a hot spot in the same lung area as the x-ray,” said Dr. Moores. “When we fused that together with the results of V/Q scanning, we saw an area of Tc99m apcitide accumulation for the clot, an area of decreased perfusion distal identifying an acute clot with a distal perfusion defect.”

Currently the FDA-approved indications for Tc99m apcitide are equivocal or negative US studies with a high degree of clinical suspicion of DVT. It is also labeled for patients with a prior history of DVT presenting with recurrent symptoms.

Talliefer looked at the use of Tc99m apcitide for diagnosing DVT. Thirty-nine patients underwent Tc99m studies and venography. Images were taken at 10, 60 and 120 minutes. Sensitivity improved over time rising from 63.6% specificity and 71.8% accuracy at 10 minutes, to 76.2% and 81.6% respectively at 120 minutes, and 86.4 and 87.2% respectively when all tests were reviewed together (Taillefer R, et al. J Nucl Med. 1999;40:2029).

A follow-up study looked at overall accuracy of Tc99m apcitide in DVT diagnosis. The sensitivity of Tc99m apcitide compared to contrast venography was 74% when read centrally by expert readers and 81% when read at the local institution. In a subset of patients with no prior history of DVT, the sensitivity rose to almost 100% locally. The negative predictive value at varying prevalences of 20-50% ranged from 90-97% (Taillefer, R et al. J Nucl Med 2000;41: 1214).

“Tc99m apcitide hasn’t been formally studied for PE,” said Dr. Moores. “There may be a place for these studies in obese patients, trauma victims with casts, isolated calf disease, high-risk patients with equivocal US findings and other instances of problematic imaging.”

Evolution of Approaches to the Solitary Pulmonary Nodule

Thaddeus Bartter, MD, FCCP, from the University of Medicine and Dentistry of New Jersey in Camden, noted that, “The study of the solitary pulmonary nodule (SPN) will always be crucial in pulmonary medicine.”

The central problem with SPNs, according to Dr. Bartter, is the issue of malignancy. If malignant, it has been discovered at an early stage and most undergoing resection will be cured.

However, a resectional approach to all SPNs is not tenable as a substantial percentage of SPNs are benign. To subject all patients with SPN to surgery, or even invasive testing, will inevitably cause morbidity and mortality in patients who would have remained well if left alone.

Dr. Bartter simplified the extensive differential diagnosis in SPN suggesting that the differential be grouped into three major concerns: cancer, tuberculosis (and other granulomatous infection), and “other.”

Each category calls for a different type of intervention. The dilemma for the clinician is to be able to separate the three.

Dr. Bartter reviewed the basic definition of SPN, a rounded discrete parenchymal density surrounded by lung tissue. He pointed out that the traditional size range used in most of the literature is 5-30 mm.

“I think that the name ‘SPN’ needs to be modified,” said Dr. Bartter. “The correct term should be ‘CT-indeterminate SPN’.”

CT scanning has had a major impact upon the clinical approach to SPN. Since the advent of CT scanning, many more SPNs are identified than in the days when plain films were the only available modality.

A patient with a normal plain film may have an SPN on CT, and a patient with a single nodule on plain film may be found, on CT, to have multiple lesions. CT scanning, with its detailed information, has allowed us to determine that many densities thought to be SPNs on plain film are almost certainly benign.

In discussing what percent of SPNs are malignant, an article by Swensen was reviewed. He looked at 629 SPNs 4 to 30 mm in size. Of these 60% were benign, 23% were malignant and 12% non-diagnosed (Swensen SJ, et al. Arch Intern Med. 1997;157:849). The percentage that is malignant varies significantly between studies, but is significant in all of them.

Dr. Bartter reviewed the available diagnostic tools and postulated three possible approaches to an SPN: “the intellectual biopsy,” “tissue biopsy,” and “biological biopsy”.

An “intellectual biopsy” (IB) is compiling facts to determine risk of malignancy taking into account factors such as size, the age of the patient, and smoking history. IB describes a formalized statistical approach to this data.

An example is the Bayesian model used by Cummings for SPN. For each of the factors above, different likelihood ratios (LR) were assigned (for example, a low likelihood ratio for a non-smoker and a higher one for a 30-pack-year smoker). The LRs are then multiplied by each other to yield a statistical risk of malignancy. The implication is that for a very low-risk lesion, observation would be the safest approach, whereas a very high-risk lesion should probably be operated upon (Cummings SR, et al. ARRD. 1986;134: 449).

This statistical process does not appear to offer a benefit over the analysis of a treating physician. Swensen in 1999 compared a mathematically driven model similar to Cummings’ with the ability of different physicians — a general internist, a thoracic surgeon and a chest radiologist — to predict likelihood of cancer. He found that the physicians’ capacity to predict malignancy were identical to the mathematical model (Swensen SJ, et al. Mayo Clin Proc. 1999;74:319).

“Our clinical judgment is just as good as the formalized computer programs,” said Dr. Bartter. “There is no advantage to these systems except to verify judgments against an external standard.”

The meaning of “tissue biopsy” is self-evident. This approach is not without its problems, the largest being diagnostic accuracy. A bronchoscopic biopsy or fine needle aspiration biopsy, if positive for malignancy, would lead to a surgical recommendation. If either biopsy is negative, however, one cannot be certain that the lesion is not a malignancy missed by the biopsy. In a high-risk patient the recommendation would still be surgery, making the procedure almost pointless.

The accuracy of tissue biopsies increases with thoracoscopy, which is nearly 100% diagnostic and has equal sensitivity for benign and malignant diagnoses. The issues with thoracoscopy are finding the lesion and the invasiveness of the procedure. With low mortality and morbidity it is a good tool for diagnosing accessible nodules.

“Biological biopsy” is a term Dr. Bartter used to describe methods which identify biological characteristics of SPNs in an attempt to separate malignant from benign. Dr. Bartter discussed dynamic CT, Positron Emission Tomography (PET) scanning, and Tc99m depreotide scanning.

Dynamic CT involves rapid serial imaging of an SPN after contrast injection. Cancers, because of their vascularity, will enhance after injection. Swensen published results of a multicenter trial using dynamic CT showing that malignant lesions usually enhanced to a greater extent than benign ones. Using an enhancement of >15 Houndsfield Units as cutoff, dynamic CT had a 98% specificity for cancer if enhancement occurred and a 96% negative predictive value if there was no enhancement (Swensen SJ, et al. Radiology. 2000;214:73).

PET scanning and Tc99m depreotide scanning identify different biological properties of cells present in higher concentration in malignancies than in most other lesions. With both studies, a malignancy “lights up.”

Dewan showed that PET has significantly better diagnostic accuracy for an SPN than Bayesian analysis (Dewan NA, et al. Chest. 1997;112:416). The results of a negative or positive PET study might, therefore, have a significant impact upon a wait-and-watch versus an operate-now approach.

Tc99m depreotide has the same potential. In a study comparing PET and Tc99m depreotide, Blum et al. found the sensitivity and accuracy of the two methods were essentially identical (Blum, et al. Chest. 2000;117:1232).

Dr. Bartter then moved to issues of staging early lung cancers. In patients with a malignant SPN, which is surgically resected, one in three will be dead in five years. The inevitable conclusion is that tumor cells are escaping. The mediastinal nodes are the usual portals for dissemination.

There may be a significant difference between treating a positive mediastinum after it has been discovered during resection and treating it pre-operatively with chemotherapy (neoadjuvant therapy). In key phase III trials, neoadjuvant therapy was shown to significantly increase survival in patients with stage III non-small-cell lung cancers (NSCC) (Rosell R, et al. NEJM. 1994; 330:153; Roth JA, et al. J Natl Cancer Inst. 1994;86:673; Depierre A, et al. J Clin Oncol. 2002;20:247). Therefore, accurate mediastinal staging may have an impact upon treatment decisions and ultimate survival.

Webb completed a retrospective study of 170 patients with known or suspected NSCC randomized to CT or MRI for staging. For both nodal stations sensitivity and specificity were, in Dr Bartter’s view, “dismal” (Webb RW, et al. Radiology 1991;178:705).

“I would never allow my mother to be treated with a modality that has this little sensitivity and specificity,” he stated. “Probably because nothing else was available, we’ve been using this poor data for years.”

Mediastinal staging using biological imaging, PET or Tc99m depreotide, is significantly more accurate than CT staging. Dowenka completed a meta-analysis comparing CT and PET scanning. PET scanning had significantly better accuracy, positive predictive value and negative predictive value than CT (all P<.001) (Dowenka BA, et al. Radiology. 1992;213:530).

With respect to Tc99m depreotide, an abstract by Waxman et al. showed that staging was 81% sensitive and 86% specific with a negative predictive value of 98% and a positive predictive value of 41% (Waxman AD, et al. SNM 2002, abstract).

“Two of the false negatives were in areas adjacent to primary tumors, and no nuclear medicine study can see a flashlight behind a lighthouse,” said Dr. Bartter. “Also, ‘false positive’ nodes were mainly in the drainage of the tumor and may have been a result of sampling error on the part of the pathologist.”

This raised the intriguing possibility that the “false positive” nodes may actually have been true positives picked up by Tc99m depreotide scanning but missed on pathologic sampling.

Overall, Dr. Bartter believes biological imaging is the most accurate non-surgical tool available to identify lung cancers. Tc99m depreotide or PET studies may soon become a routine approach both to the analysis of SPNs and staging of primary lung cancer.

Tc99m Depreotide in the Diagnosis of Early Lung Cancers

“The whole idea of functional imaging is to change what you are going to do with your patient,” said Robert S. Bridwell, MD, from the Uniformed Services University of the Health Sciences, Bethesda, MD. “If there is no impact on clinical management, then don’t order the test. If you are going to make decisions based on a positive or negative study, these are the patients you want to use functional imaging on.”

Functional imaging (FI) may be most useful in cases where radiographs are inconclusive. A positive functional study helps the physician risk-stratify the patient to aggressive or non-aggressive diagnostic algorithms.

Using dual time-point Tc99m depreotide imaging studies may help lessen positive readings. In infections, there may be amorphous uptake early. However, over time the signal becomes less intense, as the lesion washes out. It is, as Dr. Bridwell says, “behaving well”.

“Adding an extra sequence is how we improve the specificity of both PET and depreotide imaging,” he noted. “If I have a positive scan at two hours, I will often bring those patients back at four hours to minimize false positives.”

Another use is establishing a functional road map. Depreotide is injected into the patient 18 to 24 hours prior to surgery. Using an intra-operative gamma probe, the surgeon finds the hottest tissue. This information can be used to point pathologists toward nodes that may harbor micrometastic disease.

“An interesting new concept we are studying is applying the techniques of sentinel node breast cancer to the lung cancer patient,” said Dr. Bridwell. “Using Tc99m depreotide as our probe, we try to find tissue likely to harbor micrometastatic disease.”

PET scans have many false positives with sarcoid and infections. He highlighted one patient with sarcoid and new-onset right bundle branch block. After administering Tc99m depreotide, uptake was seen in the left ventricle (LV). This pattern was never seen in cancer patients studied earlier.

Based on an over-expression of somatostatin receptors in the LV, consistent with acute inflammatory disease, the patient was given steroids. About halfway through the course, Dr. Bridwell repeated the study and found no abnormal accumulation of radiopharmaceutical in the LV.

“Chest x-rays throughout the course were unchanged,” said Dr. Brid-well. “Functionally, we can see the steroids are treating the acute disease.”

Polypeptide platforms are being studied to deliver radiotherapy, similar to radioactive iodine use in thyroid cancers. Researchers are studying the possibility of removing the Tc99m from depreotide and adding a high-energy beta emitter to deliver therapy to the tumor.

The Periodic Table of the Elements shows that Tc99m is underneath rhenium 188, a high-energy beta emitter. This means that the two have the same chemistry. Phase I trials show it delivered high levels of radiation to 80% of the tumors with no effect on kidneys, liver, spleen, or bone marrow.

“If the trials continue successfully, this may be the future of polypeptide platforms,” said Dr. Bridwell. “There is no immunogenicity associated with the polypeptides. So, unlike the antibody platforms, reinjection is an option.”