A 40-year-old wife and mother, whom I’ll call Nancy, complained of dizziness an hour or so after eating. Her family physician documented low blood sugar, hypoglycemia, but her fasting insulin levels were normal, as was a CT scan of her pancreas. Her symptoms continued without a diagnosis. A year later, she developed pain in her upper abdomen and was found to have gallstones. She was referred to me for removal of her gallbladder, a cholecystectomy. Her symptoms didn’t fit the classic pattern of tumor-induced hypoglycemia but nonetheless suggested that her pancreas, which produces several hormones which regulate blood sugar levels, was behaving abnormally. We agreed that during her cholecystectomy, I would carefully inspect her pancreas for signs of an insulin-secreting tumor or insulinoma, or any other abnormality.
The operation was performed through a midline abdominal incision as I cared for Nancy a few years before laparoscopic cholecystectomy was introduced. After the gallbladder was removed, the pancreas was exposed so I could look at and feel every part of it. The pancreas appeared completely normal, a pale beige-yellow gland full of soft lumps or lobules, about 10 inches long and 2 inches across, resting behind the stomach and draped across the major blood vessels in the abdomen (Figure 1).
Figure 1. The Pancreas location within the abdomen. Inset: The areas of the pancreas with the adjacent first portion of the small intestine or duodenum and the bile duct that passes through the pancreatic head.
The “head” or thickest part of the gland is joined to the first portion of the intestine or duodenum. It then narrows to the “neck” which straddles across major blood vessels and continues through the center of the gland, which is referred to as the “body”, which then tapers to the “tail” which extends all the way to the spleen on the left side of the patient’s abdomen. I gently squeezed every part of the gland between two fingers, feeling for any hard lumps or “kernels” which might be evidence of a tumor. Nothing felt abnormal.
Nancy recovered uneventfully from the surgery. I told her to try to follow a low-carbohydrate diet, as reactive hypoglycemia, which can occur after eating, is usually improved by this. I didn’t hear any more from her until three years later, after she began losing weight and felt ill. A new CT scan showed an irregular, lemon-sized mass in the tail of the pancreas and multiple dark spots in her liver which were typical of cancer spread, or metastases. A needle biopsy confirmed metastatic pancreatic cancer. She was offered chemotherapy, but there was little chance that anything would halt the growth of her cancer.
She returned to see me accompanied by her daughter, now eight years old, who had long blond curls just like her mom. “Why didn’t you find the cancer when you operated on me?” Nancy asked. Tears welled up in her eyes, and her daughter stood close behind her, silently staring at me with a look of worry on her face. Nancy died within six months.
Pancreatic cancer is an insidious disease that often eludes diagnosis until the tumor has spread to other organs and tissues and is then incurable. This is due to the location of the pancreas deep in the abdomen, where it cannot be felt on physical exam, and to the fact that a tumor of this fleshy gland can grow large before it produces specific symptoms. In Nancy’s case, despite my suspicion that a pancreatic abnormality was present, and despite a hands-on inspection of the pancreas, her early cancer, which almost certainly had been present, was undetectable. I was frustrated and saddened by this experience but had to accept that, given the then-current state of available treatments, this sequence of events was all too common for this illness.
The Challenge of Early Detection
We know that the common form of pancreatic cancer, called pancreatic ductal adenocarcinoma, or PDAC, has a long “latency period,” in which it develops and grows slowly and silently at first, but then accelerates and spreads rapidly. From studies on cancer cell growth, we believe that it can take up to 10 years for a cancer to grow from a microscopic or very early stage (called carcinoma-in-situ) to the point that it is apparent on a CT scan. Tumors which are no larger than 2 cm (about ¾ of an inch) are called Stage I tumors, but even Stage I tumors are found to have already spread beyond the pancreas more than half the time, and they are frequently missed on CT scan.
When CT scanning first became widely available in the 1970s, we were able to see masses – tumors – of the pancreas that had previously been hidden. Prior to that time, the presence of a mass or tumor of the pancreas was apparent on x-ray only if it distorted or blocked part of the GI tract made visible after the patient drank barium and could be seen on the x-ray. These tumors were almost always large, either too large to remove or too large to eliminate by surgically removing them. Often, the tumors occurred in the pancreatic head and pinched off, or blocked, the common bile duct, through which bile – the syrupy liquid that is produced by the liver – flows into the duodenum. This blockage produced jaundice – the yellow discoloration of eyes and skin – that led to the discovery of the tumor. But when the tumor occurred further away from the area of the bile duct, we usually couldn’t detect it until it was large or already widespread.
Only about 6% of PDACs are discovered when the tumor is classified as Stage I. Not coincidentally, this is also the approximate “cure rate” of this cancer. The problem is that although CT scanning is the most common method of detection of PDAC, worldwide, it frequently fails to detect the Stage I tumor. There simply isn’t enough of a difference in the CT scan imagery of a normal pancreas and a pancreas with a small cancerous tumor. Magnetic resonance imaging, or MRI, is better for detecting these small cancers, but MRI is less available and more expensive than CT scanning. An unfortunate reality is that the most common method to look for PDAC, the CT scan, is frequently unable to detect it when it is still curable.
Within a few years after Nancy’s surgery, ultrasound was being used to find small tumors of the pancreas. Ultrasound is a technique where sound waves are transmitted from a transducer, and the reflections of the sound waves are converted into images which reveal the characteristics of the tissue being scanned. At the time of Nancy’s illness, ultrasound was being used to detect gallstones when the transducer was aimed at the area of the gallbladder or to examine the growing fetus in a pregnant uterus. Abdominal ultrasound exams usually didn’t give us very good views of the pancreas, however, because of acoustical interference by the overlying bowel. But we subsequently discovered that if we applied the ultrasound transducer directly over the pancreas during surgery, we could see the pancreatic anatomy in exquisite detail. When the transducer was held right next to the exposed pancreas, abnormalities which were only 1 or 2 millimeters in size were easily visible (Figure 2). This was a major discovery but was only useful if we were already operating on a patient with a known abnormality of the pancreas.
Figure 2. Intraoperative ultrasound of the head of the pancreas showing a tumor (dark arrow) near the duodenum (white arrow).
Beginning in the 1980s, ultrasound transducers were miniaturized and were mounted on the end of flexible endoscopes that could be passed through the mouth and stomach, where the pancreas could be viewed or imaged by aiming the transducer through the back wall of the stomach or duodenum (Figure 3). Endoscopic ultrasound, or EUS, has revolutionized our ability to detect small pancreatic tumors that cannot be seen on CT scan or felt by the surgeon. If it had been available at the time of Nancy’s surgery, we might have been able to detect her early pancreatic cancer.
Figure 3. Endoscopic ultrasound (“EUS”) image of the left, and artist’s depiction of the same image on the right, showing the dark tumor measuring just one centimeter in diameter just below the circular head of the endoscope
EUS is an endoscopic procedure performed in an ambulatory center or hospital, in which the patient is briefly sedated or anesthetized, and the endoscopy is performed by a specially trained gastroenterologist. It is invasive – the patient has to undergo the procedure of having the endoscope advanced to the area where the tumor can be seen – and it is highly dependent on the skill of the endoscopist. It is not used as a general screening test and isn’t as widely available as either CT scanning or MRI. It is usually only recommended when the CT or MRI looks abnormal. Bioengineers and radiologists have been working to develop better, non-invasive ways to detect small cancers. Artificial intelligence methods, which include deep learning of computerized images, have the potential of achieving the same fidelity as EUS. Clinical trials of these new methods are currently underway.
A major problem with PDAC is that there is no good blood test to tell us when an early pancreatic cancer is forming. Such a test – called a biomarker – is desperately needed for this disease. Breast cancer can be detected at an early stage by physical exam or by mammography. Prostate cancer can be detected by physical exam, or with a blood test which measures the amount of prostate-specific antigen, or PSA. Lung cancers can be detected on a simple chest x-ray, and colon cancers, or their pre-cancerous growths, can be found by means of a colonoscopy. No such simple examination or “non-invasive” test exists for the early detection of pancreatic cancer.
The search for biomarkers of cancer led investigators in the 1980s to discover that many cancers produce proteins that are normally produced during the development of the embryo during pregnancy, the production of which is then shut down after birth. These embryonic proteins are produced abnormally, along with many other growth factors by the cancer, and specific assays for these proteins, called onco-fetal antigens, indicate the presence of some tumors. Colon cancers can be detected by a rise in the level of a protein called carcino-embryonic antigen, or CEA. Ovarian cancer can be detected by a rise in a protein called CA-125, and pancreatic cancer produces a protein called CA19-9. Serum levels of CA19-9 don’t usually rise, however, until the tumor is already large and metastatic. So, the level of CA19-9 circulating in the bloodstream is not a reliable test for early-stage pancreatic cancer.
The search for an accurate and sensitive biomarker of pancreatic cancer has intensified over the past decade. Proteins, combinations of proteins, and fragments of RNA – ribonucleic acid produced by the cancer cells to direct the production of proteins – have been examined to discover a marker of early-stage pancreatic cancer. There are now hundreds of marker candidates, but the tests to verify their usefulness are complex (and expensive), and the problem is that so few early-stage cancers are found that a study of the biomarkers will require hundreds, or even thousands, of patients to obtain enough early-stage patients to confirm the value of the test.
Most recently, there has been intense interest in surveying a patient’s blood for characteristic patterns of proteins that are produced by the body in response to the presence of the tumor, rather than those which are produced by the tumor itself. This shift in focus has attracted much attention, and large-scale studies to validate these tests are now underway.
Although the number of people with lung, breast, colon, and prostate cancer is much greater than those with pancreatic cancer, the five-year mortality or death rate of patients with pancreas cancer is about 92%, which is much higher than for all other malignancies. Currently, there are about 245,000 invasive breast cancers diagnosed each year in the United States, and the five-year survival has risen to more than 80%. So, more people now die of pancreatic cancer than die from breast cancer. The number of patients dying of pancreas cancer exceeds those who die of all of the other common cancers with the exception of colon cancer (which kills slightly more people than pancreas cancer) and lung cancer. The death rates of all of these other cancers have been steadily falling over the past few decades, however, due to better surveillance and screening programs and more effective prevention and therapy. But the deaths due to pancreas cancer continue to climb, and it is estimated that by 2050, pancreas cancer will become the leading cause of cancer-related death in this country. So far, nothing has changed this terrible trajectory.