The publicity material for M. D. Anderson in Houston, Texas, which is perhaps the leading cancer treatment center in the United States, frequently includes a statement to the effect that they are committed to “ending cancer.”
That statement puts me in mind of maps dating from the first century AD, in which large unexplored areas were tagged with the phrase “hic leones sunt.” Meaning, “here there be lions.” Or, sometimes “hic dragones sunt.” Even when the Roman Empire extended to the eastern shores of the Mediterranean, including Egypt and Palestine, there were vast stretches of the planet about which the map makers had not the slightest clue, so they tagged the unknown territory with those phrases, indicating unknown perils.
“Cancer” as a generic phrase has some of that same connotation – a large scary area filled with dire threats. But the cancer map has been filled in to a very great degree. The National Cancer Institute has a list of cancer types that runs to about 200 items. Granted, there’s a certain amount of overlap. For instance, Kaposi Sarcoma is listed as an AIDS-related cancer and also has its own separate listing. However, breast cancer gets a single listing; the several distinct types of breast cancer such as HER-2 and triple-negative breast cancers do not get separate listings. And while the Mayo Clinic lists vestibular Schwannoma as a type of acoustic neuroma, the NCI list includes neither. I did not do an item-by-item comparison of those two lists, and there are other lists besides those two. But, as you can see, the number of distinct cancer types grows.
Evidently, the cancer map has become highly detailed, like one of those atlases that show every little village and all the interconnecting roads. And that’s what you need to really explore the country. If all you do is whiz by on the interstate, you’re missing a great deal. That’s what cancer treatment was like when all that could be done is attack “cancer” as a single entity.
No more. For example, on the roads in our town there are little signs in front of people’s houses with the exhortation “Beat DIPG” under the name of a young child, to which is affixed the motto “We Believe.”
DIPG stands for diffuse intrinsic pontine glioma, which I had never heard of until I did some sleuthing. Gliomas are brain tumors; pontine means that these tumors are located in the pons, which is part of the brain stem. Intrinsic means that the tumors generally stay within the pons, but diffuse means that the tumor cells travel within the pons and sometimes into the spinal column, which is below the pons. You may remember that what brought Senator McCain to his end was a glioma. However, diffuse intrinsic pontine gliomas mostly affect children, including quite young children. The child in our town who was diagnosed with DIPG is eight years old. As things now stand, her chances of surviving longer than two years are less than 10%, and her chances of five-year survival are less than 1%.
Nevertheless, the “Beat DIPG” slogan is not howling at the moon. The Weill-Cornell Brain and Spine Center has established a Children’s Brain Tumor project, and clinical trials are under way in an effort to at least identify promising lines of research. Several other research centers are investigating combination treatment using chemotherapy in combination with radiotherapy, and the National Institute of Neurological Disorders and Stroke, part of NIH, is conducting a trial of an experimental drug for the treatment of DIPG. Currently, there are 69 clinical trials, recruiting or underway, in DIPG. Some of the drugs being investigated in these trials include gemcitabine, panobinostat in a nanoparticle formulation, erlotinib, everolimus, dasatinab, mephalan HCl, temsirolimus, vorinostat, bevacizumab, and a bunch of others. Institutions behind these trials include Memorial Sloan-Kettering, the aforementioned M.D. Anderson, the Mayo Clinic, and academic centers in the US and Europe.
A major obstacle to research is the scarcity of tumor tissue samples; an attempt to address this is being carried out by the Centers for Genetic Medicine. Whether any hopeful treatment forms will emerge in time to change the outcome for the eight-year old girl in our town is doubtful. However, that little girl is enrolled in a clinical trial. I do not know the particulars of that clinical trial, but I certainly hope for success in the trial and the best possible outcome for that child.
DIPG is, fortunately, relatively rare. About 200 to 300 children in the US are diagnosed with DIPG every year. By the way, neither the NIH nor the Mayo Clinic lists of cancer types include either DIPG or gliomas; these would all be subcategories under brain cancer.
So, while M. D. Anderson continues to sound the “End Cancer” bugle call, the targets have become ever more specific. As we’ll see when we take a look at the most recent developments on the cancer front, narrowly targeting specific cancers not only potentially benefits patients with those cancers, but also greatly benefits the pharmaceutical companies that originate and develop those new drugs. The rationale is simple and logical. It’s much easier to get regulatory approval for a drug that treats a specific disease for which there is no other treatment than for a drug that addresses a broader disease category, for which some treatment options already exist. If there are no other treatment options, and your candidate drug confers at least some benefit, you are likely to get your drug approved. But if there are other treatment options, you need to demonstrate that your new drug is at a minimum equivalent to the existing drugs, and also that your new drug is in some way better than the existing drugs. Superior efficacy would be the big winner, but a better side effects profile, or more convenient dosing, or a combination of other factors, could also score points towards regulatory approval.
So, let’s say that one of those clinical trials in DIPG turns up a mode of treatment that delivers some benefit. Maybe it increases the mean five-year survival from the current less than 1% of patients to perhaps 35% of patients, which was the five-year survival rate for all brain cancers during the years from 2008 – 2014. That’s a huge benefit for those patients, but as we’ve learned, patients with DIPG are not numerous – just a few hundred per year. Now the drug maker takes aim at a larger cohort of patients – the totality of patients with any form of brain cancer. That number is considerably larger – the estimated number of new cases in 2019 is 23,820, and the estimated number of deaths is 17,760. If, as we said earlier, that candidate drug was shown to deliver some benefit to brain cancer patients, it would be likely to get regulatory approval.
While we’re discussing matters such as the estimated number of new cases of various cancers and the mean five-year survival for those cancers, here’s an updated table that first appeared in a piece that posted of October 6, 2018. The number of estimated cases haves been updated to the year 2019; the estimated percentages of patients who will still be alive five years after diagnosis have improved slightly.
|Type of cancer||Estimated New Cases in Males – No. & % of Total||Estimated New Cases in Females – No. & % of Total||Estimated 5-year Survival – M & F|
|Lung & bronchus||116,440 14%||111,760 12%||20%|
|Colon & rectum||78,500 9%||64,010 8%||66%|
|Urinary bladder||61,700 7%||18,770||78%|
|Melanoma||57,220 6%||39,260 5%||94%|
|Kidney & pelvis||44,120 5%||29,700 3%||75%|
|Non-Hodgkin lymphoma||41,090 5%||33,110 4%||74%|
|Leukemia||35,920 4%||25,860 3%||65%|
|Oral & pharynx||38,140 4%||14,860 2%||68%|
|Liver & bile duct||29,480 3%||12,550 2%||19%|
|Pancreas||29,940 3%||26,830 3%||9%|
|All cancer sites||870,970||891,480||69%|
The number of new cancer cases continues to rise, and will almost certainly go on rising. This is because there has been, and will continue to be, a large decline in heart disease in the US and the rest of the developed world. If heart disease rates had remained at their 1996 peak, there would have been about 10 million more deaths attributable to heart disease since then. Since, inevitably, those 10 million individuals will take leave of life on Planet Earth, the likelihood is that the mechanism that carries them off will be cancer.
Nonetheless, the rate of cancer deaths in the US has declined considerably. If the overall cancer death rate in 1990 had persisted, the result would have been about 2.4 million more cancer deaths between 1991 and 2015. Cancer deaths are projected to continue to decline, despite the increase in the number of cancer diagnoses. Overall, age-adjusted cancer mortality in the US declined by 26% from 1991 to 2015.
The decline in cancer mortality is clearly shown in the increase in patients who survive for five years after diagnosis. Overall, the five-year survival rate has jumped by 20 percentage points, from 49% in 1975 to 69% by 2014. For some individual cancers, the improvements in this marker are impressive: for prostate cancer, from 68% to 99%; for breast cancer, from 75% to 91%; for stomach cancer, from 15% to 32%; for liver cancer, from 3% to 19%; and even for pancreatic cancer, perhaps the most dire cancer diagnosis, from 3% to 9%. (I wish there were equivalent data for 10, 20, and even longer survival intervals. My guess is that a high proportion of five-year survivors just “keep on keeping on.”)
However, it should be noted that those five-year survival rates should not be taken as cure rates. A man who has is still alive five years after a diagnosis of prostate cancer has not necessarily been cured. Perhaps the cancer is merely indolent. These rates are partly indicative of the aggressiveness of a particular form of cancer, as well as being partly indicative of the likelihood of effectiveness of treatment for that cancer. For example, the five-year survival rate for prostate cancer combines data about the men who have opted for definitive treatment, such as prostatectomy, with data about the men who have opted for “watchful waiting.”
It’s hardly a coincidence that the cancers with the best five-year survival rates are the ones that are most susceptible to early detection – breast, prostate, melanoma of the skin, thyroid cancer – while pancreatic and liver cancers often are detected when they have already progressed to the point where treatment options are inadequate. The emphasis on early detection and early intervention pays off.
The means of detection vary greatly, depending on the location of the cancer, and the forms the cancer takes. Solid tumors growing in places where they can be palpated or X-rayed can be detected fairly easily. The deeper inside the body, the more difficult it is to identify a cancerous growth. The procedures employed to try to detect cancers in various parts of the body may be invasive, are frequently troublesome to the patient, may in some cases present a degree of risk to the patient, and are not uniformly accurate. The limitations of these tests are fairly well publicized, with the result that many persons avoid these tests. Another factor is the position of the U. S. Preventive Services Task Force (USPSTF) on both the prostate-specific antigen (PSA) test for prostate cancer and on mammograms. These are well known and have led to a reduction in the proportions of men having the PSA test and women having mammograms, with the result that more of these cancers are now being initially diagnosed at a more advanced stage.
Colon cancer is the third highest cancer cause of death. Colonoscopies are about 95% effective in finding and removing the polyps that would grow into cancers. The established wisdom is that everyone should have colonoscopies starting at about age 50, and then at intervals ranging from three years up to seven years, depending on findings and on the patient’s age. Yet, only a bit more than half of the US population in the 50 to 65 age bracket has ever had a colonoscopy. That proportion rises slightly, to just about two-thirds of those of us over the age of 65.
As we saw from the table above, about one-third of all patients diagnosed with colon cancer die within five years from diagnosis. Given the high degree of effectiveness of colonoscopy in preventing the development of full-blown colon cancer, it is evident that the great majority of those deaths are in patients who have not had a colonoscopy. The inconvenience and, yes, the unpleasantness of a colonoscopy is undeniable (I have had four!) but equally undeniable is that a colonoscopy is far preferable to the alternative.
The ideal would be a single biomarker for all cancers… or would it? More than one such biomarker has been proclaimed to be forthcoming in recent years, frequently linked with a campaign to get folks to invest in a biotech outfit that will make them billionaires. (Remember Theranos and Elizabeth Holmes, soon to go on trial?) So far, no dice. But how useful would such a biomarker be? Supposing you went for your annual physical, and your physician reported, with a worried look, that your universal cancer test had come back positive. Which is to say, “you have cancer somewhere in your body.” Somewhere, but where?
In the opinion of your skeptical Doc Gumshoe, such a test, if there ever was one, would bring your physician no closer to a genuine treatment plan than when you walked in the door for your physical. And, I would add, it’s highly likely that there would be a huge proportion of false positives with such a test – that is to say, the test would report the presence of cancers that would never progress to the stage of requiring treatment. That’s because cancers emerge in our bodies all the time, as a result of random errors in genetic transcription (mutations) as our cells multiply. Most of those errors are meaningless and will not be carried forward. But a few those errors may eliminate the cell characteristic that ensures that cells have a limited lifetime. A limited lifetime for cells, or programmed cell death (apoptosis)– is beneficial, since it permits evolution – once in a while a transcription error results in a beneficial quality. Those cells without that programmed death characteristic are where cancer starts and spreads. Fortunately – and that’s an understatement – our worthy immune system identifies most of those cancerous cells fairly early on and wipes them out. It is the relatively few that escape the immune system that progress to become cancers
But that single cancer biomarker might identify those cancers that aren’t going anywhere along with the truly menacing cancers, with the result that the much-touted test for the all cancers might end up pointing to nothing of value.
That doesn’t mean that the search for biomarkers is in vain, or hasn’t already produced some valuable results. One method, for example, has produced quite promising results in several different cancers.
Grail’s liquid biopsy may be able to detect multiple cancers
Grail, a start-up in Menlo Park, California, has received FDA breakthrough designation for its liquid biopsy technique. Grail is a spin-off from the DNA sequencing company Illumina (ILMN), and the technique combines Illumina’s ultra-deep DNA sequencing, which reads a DNA region about 50,000 times, with a machine learning algorithm developed by Grail, which identifies mutations. The method was tested at Memorial Sloan-Kettering against a group of 127 patients with advanced metastatic non-small-cell lung cancer (NSCLC). In a subgroup of 91 patients whose cancer-driving mutations had been identified by means of tissue biopsies, the liquid biopsy identified 68, for a positive rate of 75%. There were also 19 patients without mutations identified through tissue biopsies. In these patients, the liquid biopsies detected no irregularities, meaning that the false positive rate with this method was zero.
Grail also presented data that their liquid biopsy method achieved detection rates ranging from 59% to 92% in patients with adenocarcioma, squamous cell, and small cell lung cancers, while at the same time maintaining a false positive rate of less than 2%.
At the American Society of Clinical Oncology (ASCO) meeting this year, Grail presented data from its Circulating Cell-free Genome Atlas (CCGA) project, reporting detection rates in a dozen prespecified and potentially fatal types of cancers, in earlier stages. Sensitivity rates for stage I – III cancers were as follows: anorectal (79%), colorectal (74%), esophageal (76%), gastric (78%), head and neck (86%), hormone receptor negative breast (64%), liver (68%), lung (59%), ovarian (67%) and pancreatic (78%) tumors, as well as multiple myeloma (71%) and lymphomas (70%), excluding leukemias. Sensitivity by cancer stages was: Stage I, 34%; Stage II, 77%, Stage III, 84%, and Stage IV, 92%.
Grail has further studies under way. The STRIVE Study will track more than 100,000 women undergoing mammograms to validate the ability of its liquid biopsy test to detect breast cancer and other malignancies. Results are due to report in 2020. Another study, SUMMIT, will enroll about 50,000 persons in London who have not been diagnosed with cancer, but are thought to be at high risk for lung cancer due to smoking histories.
Grail’s approach to detecting cancer is to focus on DNA methylation, a process used by cells to regulate gene expression. In cancer, abnormal methylation patterns and changes in expression can contribute to tumor growth, including the suppression of tumor-suppressing genes.
Doc Gumshoe agrees that Grail’s results so far are definitely promising. However, some perspective is needed. A 92% sensitivity rate for Stage IV cancer in all probability only confirms what we knew already. On the other hand, a 34% sensitivity rate for Stage I cancer essentially means that two out of three patients with Stage I cancers escape detection, at least until their cancer progresses. However, adding liquid biopsy to the options for detecting cancer is a highly significant step in the desired direction.
A positive response to a liquid biopsy test would almost inevitably lead to imaging studies to pinpoint the location of the tumor, determine the stage of the cancer, whether the cancer had metastasized, and the location of metastases. Various types of imaging studies can be used, including magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT). Stuart Taylor, MD, of the Centre for Medical Imaging at University College, London, noted that “the accuracies of these various scanning tests differ between the different organs. Because the scans differ in their ability to find disease, patients often undergo a whole range of scans before their tumor is fully staged and the treatment can begin.”
Whole-body magnetic resonance imaging (WB-MRI)
The question then arises whether single whole-body magnetic resonance imaging (WB-MRI) produces results equivalent to the standard staging protocol which requires several successive imaging studies. The STREAMLINE L study was conducted at 16 English hospitals to attempt to answer this question. A cohort of 187 patients who had recently been diagnosed with non-small-cell lung cancer, most of whom had Stage II cancer or higher, was evaluated. These patients had all been through the standard staging protocol.
There were small differences between WB-MRI and the standard pathway with regard both to sensitivity and specificity. Sensitivity for WB-MRI was 50%, versus 54% for the standard pathway; specificity was 93% for WB-MRI versus 95% for the standard pathway. Thus, the standard pathway was shown to be slightly superior to the whole-body MRI protocol. However, a crucial difference between the two screening methods had to do with the time elapsed from initial diagnosis to the point where patients could begin treatment. For WB-MRI, that was a 13 day period, while for the standard pathway that period was 19 days. Thus, WB-MRI would permit patients to start receiving treatment almost one week sooner.
Patients on balance preferred the WB-MRI method to the standard pathway, although the need to spend a full hour in the MRI machine was seen by some as a distinct negative. It was also not clear whether MRI scanners with the capacity to do a whole-body scan are sufficiently widely available for this protocol to become standard practice.
In addition to the STREAMLINE L study, a similar study (STREAMLINE-C) in colorectal cancer is currently underway.
Fecal immunological tests for colorectal cancer
The standard in the US for prevention of colon cancer, as discussed earlier, is colonoscopy, but fecal immunohistochemical tests (FIT) are reported to be quite effective in detecting colorectal cancer. These are tests for fecal occult blood guided by immunologic factors specifically sensitive to blood, and these tests are able to determine with great accuracy whether there is any blood in the stool. The FIT has a reported sensitivity index greater than 90%, meaning that fewer than 10% of colorectal cancers would escape detection. Specificity is also high; fewer than 10% of results would be false positive. Note that FIT is not suggested as an alternative to colonoscopy, but as an adjunct. In Canada, FIT tests are required before insurance will approve a colonoscopy. Some authorities have expressed the hope that FIT testing will become the norm in the US as preliminary to colonoscopy, and that having both procedures available will increase the percentage of persons screened for colon cancer to 80%, which is the goal in Healthy People 2020, a program of the US Department of Health and Human Services whose goals are to attain high-quality, longer lives, free of preventable disease, disability, injury, and premature death for all people.
Combining biopsy methods improves accuracy in detecting prostate cancer
As we know by now, detecting prostate cancer starts with a simple test for prostate-specific antigen (PSA), which, for many men is a routine part of the complete blood test that they have as part of a regular physical examination. We will leave out of our discussion, at least for a moment, the lack of consensus among the various entities that have put forward their recommendations as to the ages at which men should – or should not – have PSA tests. We will jump forward, to the next step – after a man has gotten a definitely positive PSA test.
A high PSA reading does not necessarily indicate prostate cancer; there can be other reasons for elevated PSA values. Thus, a biopsy is the usual next step. There has been some debate about this; a biopsy of the prostate gland is definitely an invasive procedure, which involves pushing a hollow needle into the gland and pulling out several cores of prostate tissue. Traditionally, the physician performing the procedure just sticks the needle where he knows the prostate is, with the hope that the needle will bring out some of the tumor. This method is called cognitive targeted biopsy.
Another method of targeting the biopsy has emerged more recently – systematic transrectal ultrasound guided sampling (TRUS). A clinical trial in California, enrolling 250 men with prostate lesions detectable by MRI compared the results of each of the two targeting methods with each other, and also with the results obtained by using both methods. The cognitive targeted biopsy detection rate for significant tumors was 46.8%, while the TRUS method yielded a detection rate of 60.1%. It was calculated that using one type of biopsy alone would have missed somewhere between 11.5% and 33.3% of the 174 clinically-significant cancers detected in the cohort. The TRUS method identified lesions in 20.9% of men whose lesions had been missed by the cognitive targeted method, and the cognitive targeted method found lesions in 9.7% of men whose lesions had been missed by the TRUS method.
Clearly, neither method is perfect, but employing the two together is better than using one alone.
And one more method of attempting to determine whether a lesion in the prostate is clinically significant is multiparametric MRI, which was evaluated at the Ohio State University Wexner Medical Center. This method detected more clinically-significant cancers than ultrasound guided biopsy (38% vs. 26%), and resulted in a 13% reduction in the detection of clinically insignificant cancers. In two studies, it was reported that 28% and 27% of patients would have been able to avoid actual prostate biopsy with multiparametric MRI.
Another indicator that can be useful in predicting whether a cancer is or is not clinically significant is PSA density, which is the relationship between the PSA and the total volume of the prostate gland. Used in combination with MRI, PSA density can inform the decision whether the prostate cancer requires prompt treatment. Higher PSA density is associated with more clinically significant cancers.
A clear limitation of this method is that few urologists have access to MRI machines capable of performing multiparametric scans, and fewer still have the expertise to evaluate the results. And a question that needs to be asked is, how important is it to avoid a prostate biopsy? It’s a rapid procedure, done under a local anaesthetic, resulting in very few complications or long-term effects.
What does this mean for cancer treatment?
Particularly with regard to prostate cancer, the decision of whether to treat is greatly complicated by the knowledge that this form of cancer in particular is very slow growing, and sometimes is of little clinical consequence, unless and until it metastasizes. A benefit of the diagnostic tools described above is that they may help clinicians distinguish between more aggressive and less aggressive cancers, and thus provide treatment guidelines.
These more recent diagnostic modalities provide much more information than was formerly known about different cancers. The more we know about the antagonist, the more precisely we can target the weapons that we use to attack this antagonist. One of the first facts that was known about cancer is that cancer cells are much greedier than non-cancerous human cells. Thus, if you put a certain amount of good-tasting poison in the neighborhood of the cancer cells, it’s a good bet that the cancer cells will gobble it up a whole lot faster than the normal cells. The normal cells do suffer to some degree, but they’ll likely get over it, while the cancer cells get clobbered – at least, that’s how the thinking went.
Some cancers are still treated that way, but to a much smaller degree than what used to be the norm. The science of cancer treatment has advanced enormously, such that oncologists have a pretty good idea of each type of cancer’s greatest vulnerability. It’s like knowing that the woodchuck that sneaks into your garden in the early morning to nibble on the new green shoots is particularly fond of cantaloupe, so you use bits of cantaloupe to lure the critter into your Have-a-Heart trap.
In the next installment of the musings of Doc Gumshoe, I will survey some of the more interesting advances in cancer treatment, many of which were announced at the recent meeting of ASCO.
I look forward, as always, to receiving your comments and suggestions for further questing. Best to all, Michael Jorrin (aka Doc Gumshoe)