The Grand Illusion of Chemotherapy

A Speech to German Society of Oncology (DGO) October 28, 2000, Baden-Baden © 2000

by Ralph W. Moss, Ph.D. All Rights Reserved

It is five years since I wrote Questioning Chemotherapy. In those five years, about 275,000 peer-reviewed articles have appeared in Medline on the topic of cancer, including 25,000 on cancer chemotherapy. An additional 10,000 abstracts have been presented at the American Society of Clinical Oncology (ASCO) and the American Association for Cancer Research annual meetings. Out of this vast amount of research a total of 70 new approvals have been made by the US Food and Drug Administration (FDA) for drugs used in the treatment of cancer patients.

Has all this activity resulted in more effective treatments for cancer? Have I modified the views expressed at prior meetings of the German Society of Oncology meetings that chemotherapy is both ineffective and inappropriate for the vast majority of patients to whom it is given? Has the current crop of anticancer drugs been proven to improve the lot of cancer patients?

The brief answer is, No. If anything, the value of cancer chemotherapy is less documented today than it was five years ago. In fact, randomized clinical trials with appropriate controls are rarely performed in conventional oncology before or after the approval of a new drug. When studies are reported as positive that is usually in comparisons with other agents, or it is a measurements of surrogate markers rather than of increased survival.

Drugs may be reported to increase survival but upon closer examination this turns out to be an increase in "relapse-free survival," or "time to recurrence," and not an actual increase in median overall survival. Yet, I would argue that only an increase in overall survival conveys a true benefit to the patient. A patient who has an increased relapse-free period, but no increase in actual life time, is not actually benefited by treatment, except in a psychological sense. Real benefit in terms of overall survival is rarely demonstrated by chemotherapy for the solid tumors of adults.

This is what I call the "Grand Illusion of Chemotherapy." It is the idea that the shrinkage of tumors, or improvement in tumor markers, or increased relapse-free survival, necessarily correlate with actual benefit to patients. Surveys have shown, and common sense tells us as well, that the two outcomes that cancer patients seek from chemotherapy are (1) an improvement in quality of life, and (2) an increase in their actual survival. Tumor shrinkages are not a high priority {CA Cancer J Clin 2000;50:123-32}.

Medical oncologists, by contrast, tend to concentrate on the shrinkage of tumors. Such shrinkages are called "responses." The FDA defines a complete response as a complete disappearance of all clinical and X-ray signs of cancer for one month or more. A partial response refers to a 50 percent or greater decrease in measurable tumour size for one month or more. In the past, the FDA also required some proof of life prolongation. But this stringent requirement led to very few drug approvals. From the 1940s to the mid-1990s, in fact, only about three dozen drugs had been approved by the FDA, less than one per year. The FDA's reluctance to approve drugs based on shrinkages angered the pharmaceutical industry, as well as some oncologists and patient activists. So, in the mid-1990s the government relaxed these requirements and since then FDA has approved many new drugs. Our task is to see if these new approvals have actually resulted in improved treatments.


(in order of approval)

 Generic  Brand name  Company  Indication  Approval Date
 anastrozole Arimidex Zeneca breast IV 1996
 docetaxel Taxotere Aventis breast, lung 1996
 mitoxantrone Novantrone Immunex prostate pain 1996
 irinotecan Camptosar Pharmacia colon IV 1996
 gemcitabine Gemzar Lilly NSCLC 1996+
 topotecan Hycamptin Smith-Kline ovarian 1996
 letrozole Femara Novartis breast IV 1997
 rituximab Rituxan Genentech lymphoma 1997
 porfimer sodium Photofrin QLT lung, esophagus 1998
 capectiabine Xeloda Hoffman-LaR breast 1998++
 temozolomide Temodar Schering brain (AA) 1999

* Drugs are listed in order of initial approval. Some have had additional indications at later date.
A complete list is available at  In my discussion I have rearranged the order somewhat, but have roughly adhered to this chronological order.

+Approved for use with cisplatin
++ Approved for use with paclitaxel

I will briefly review the record of six of these drugs, plus oxaliplatin. These are the major products that have generated publicity that at times has bordered on hysteria.

SLIDE #2 Anastrozole (Arimidex) SHOWS WOMAN'S HAND IN AD

Advertisements for Arimidex (anastrozole) show a woman's hand holding a star-like tablet. It urges doctors to "put survival in the palm of her hand" and claims "56.1 percent survival" for patients treated with this drug. This figure clearly implies that more than half the women who take Arimidex are significantly benefited, if not saved from their breast cancer. However, this 56.1 percent represents two-year survival figures in studies comparing Arimidex to an older drug, Megace. In fact--as the advertisement itself notes in the fine print-women treated with Arimidex have a median time to death of 26.7 months compared to 22.5 months for patients treated with Megace. Thus, the actual difference between the two groups is 4.2 months, a difference that the study itself notes is not statistically significant {Cancer 1998;83:1142-1152}.

SLIDE #3 Docetaxel (Taxotere) AD FOR LUNG CANCER

Taxotere is a semi-synthetic derivative of the Pacific Yew tree, similar to Taxol. In non-randomized clinical trials in breast cancer, there was indeed a very high rate of responses. While only 2 out of 37 patients had complete responses, another 18 (49 percent) had partial responses. The median time to progression was 26 weeks, while with the older drug Adriamycin it was 21 weeks. Thus, at best, there was a gain of five weeks. The median survival time (which, as I have said, is the key indicator of benefit) was 15 months with Taxotere compared to 14 months with Adriamycin, a gain of one month. But 87 percent of patients had fluid retention or other serious side effects with Taxotere {J Clin Oncol 1999;17:2341-54}. One must question what the quality of life is during that extra month.

Taxotere has also been approved for the treatment of non-small cell lung cancer. According to advertisements, Taxotere is "the first single agent to show a significant one-year survival benefitwith a predictable and manageable safety profile." However, when scientists at M.D. Anderson Cancer Center compared both a high-dose and a low-dose regimen to placebo, they conceded that "overall survival was not significantly different between the three groups" {J Clin Oncol 2000;18:2354-62}. I find this study unusual in that it actually included a placebo group-something that has been eliminated from most studies performed in support of FDA approval. The failure of taxotere to impact positively on survival is omitted from the company's four-and-a-half page advertisements appearing in medical journals. There was also lung toxicity in 40 percent of those who took Taxotere as well as many other side effects.

SLIDE #4 Irinotecan (Camptosar)

In 1996, the FDA approved the Camptosar (irinotecan or CPT-11) as a treatment for advanced colon cancer. The median overall survival in one study was 10.8 months with irinotecan compared to 8.5 months with fluorouracil, a gain of two months. Overall survival was 17.4 months when irinotecan was added to the standard 5-FU-leukovorin combination, compared to 14.1 months, a gain of 3.3 months {Lancet 2000;355:1041-1047}.

Does this at least prove some modest benefit? I don't think we can conclude this, since the study in question failed to compare the drug to a placebo or best supportive care (BSC) control group. Thus, while Camptosar may be marginally better than the standard drugs, it is not necessarily better than just keeping the patient comfortable. The failure to provide a control group makes it impossible to know.

SLIDE #5: Gemcitabine (Gemzar)

In May, 1996, the FDA approved the first drug for the treatment of pancreatic cancer, Gemzar (gemcitabine). It has also been approved for the treatment of stage IV breast cancer. There was a non-randomized clinical trial at the Charité Hospital in Berlin that involved just 42 patients. It showed 6 partial responses, with an overall response rate of 14.3 percent . The median survival of patients with this treatment was 15.2 months {Anticancer Drugs 1999;10:155-62}.

There was also a 1996 British study of 40 evaluable patients, which provides an illustrative contrast. In the British study there were 3 complete and 7 partial responses. Thus the overall response rate was nearly double, at 25.0 percent. Yet in the British study the median survival was 11.5 months. or four months less than in the German study {Semin Oncol 1996;23(5 Suppl 10):77-81}. Here is yet another example of the failure of response rates to correlate in a positive way with increased overall survival.

SLIDE #6 Herceptin (Trastuzumab)

In 1998, there was great excitement over the drug Herceptin (trastuzumab). In the first clinical trial on which approval was based, 9 out of 37 patients with advanced breast cancer experienced responses after being given the new drug in combination with the standard drug, cisplatin. According to the authors of the study, "the median time to progression among the responders was 8.4 months" {Semin Oncol 1999;26S:89-95}.

In a second phase II trial, Herceptin was used as a single agent. There were 5 "responses" out of 43 patients, or about 12 percent. The minor responses in that trial lasted 5.1 months {Semin Oncol 1999;26S:78-83}. You would never known this from the publicity blitz that accompanied the drug's appearance. According to a University of California at Los Angeles news release, Herceptin represented "a significant medical breakthrough." Dr. Larry Norton of Memorial Sloan-Kettering Cancer Center in New York declared, "This is the biggest difference I have ever seen in advanced breast cancer." It is "a big effect, not a small, minor effect." This treatment "is not like anything we have ever seen before" in cancer therapy, and so forth.

In May, 1998, Genentech mounted a huge display at the ASCO meeting, announcing ahead of the FDA's deliberations that the drug would in fact soon be approved and available. This announcement created near pandemonium among breast and ovarian cancer patients, who were desperate to get the latest "cure." It made approval a certainty.

SLIDE #7: Rituximab (Rituxan)

Rituximab (Rituxan) was the first monocloncal antibody drug to receive FDA approval. This came in November 1997 for the treatment of patients with relapsed or refractory low-grade or follicular, B-cell non-Hodgkin's lymphoma (NHL). In 1998, a trial centered at M.D. Anderson Cancer Center in Houston. showed 6 percent complete responses and 42 percent partial responses with the drug. The median duration of the response was 11.6 months and the median time to progression for the responders was 13.2 months {Blood 1998;92:414a-415a}. This may indeed be a superior drug, but notice once again that it was not subjected to a randomized clinical trial comparing it to either placebo or best conventional care; overall median survival data is thus lacking.

(Incidentally, the year 2001 sales forecast for Rituxan is $549 million.)


I will also mention one drug that has not yet been approved, where in fact the FDA has resisted pressure to approve it. Oxaliplatin is an experimental drug for the treatment of colorectal cancer. By adding it to the standard 5-FU+leucovorin regimen one may increase the response rate from 16 to 53 percent, an astonishing tripling of this marker. Yet, despite this dramatic difference, the median survival was 19.4 months in the oxaliplatin-added group compared to the longer 19.9 months in the standard 5-FU+leucovorin group {J Clin Oncol 2000;18:136-47}.

Here is another illustration, out of many, where better response rates fail to correlate with increases in median overall survival. Perhaps the increased toxicity required to achieve these increased responses actually injures the patients to the point that they die sooner, thus canceling out any positive effect.

Oncologists need to abandon the Grand Illusion that response rates translate into patient benefit and accepts the fact that increased overall survival, and improved quality of life, are the legitimate goals of treatment.

Finally, I would like to say a word about the side effects of chemotherapy. These are often described as "tolerable" or "acceptable" in the medical literature. However, I have compiled a small A-to-Z guide to what chemotherapy can do the human body. (The drugs that follow in parenthesis are examples, but are hardly all-inclusive of which drugs cause which symptoms.) It is clear that anticancer drugs are among the most dangerous ever deliberately introduced into the human body for therapeutic purposes.



Aches and pains
Allergic reactions (Adriamycin, etc.)
Alopecia-hair loss and thinning (almost all agents)
Altered taste (mesna)
Anorexia (allopurinol)
Antihormonal effects (finasteride, fluatmide, leuprolide, tamoxifen, gosserelin)


Blurred vision (suramin)
Bone and joint pain (isotretinoin)


Cancer (new malignancies as long-term side effects, esp. with alkylating agents)
Cardiotoxicity (cyclophosphamide, ifosfamide, Adriamycin)
Central nervous system toxicity (methotrexate)
Chronic pain
Conjunctivitis (pentostatin)
Cystitis (cyclophosphamide and ifosfamide)


Diarrhea (ara-C)


Extreme respiratory distress (Herceptin)


Fluid retention (Taxotere)


Gallstones (octreotide)
Gynacomastia in males


Hemmorhage (cyclophosphamide)
Hepatotoxicity (chlorambucil, DTIC, topotecan)
Hyperpigmentation (bleomycin)
Hypersensitivity reactions


Increased secretion of sebaceous glands (octreotide)


Jaw pain (vincristine)


Kalemia, both hyper- and hypo-


Leukemia (as long-term side effect)


Myocardial ischemia (BCNU)


Nephrotoxicity (renal system) (cisplatin, cyclophosphamide, DTIC, ifosfamide)
Neurotoxicity (ifosfamide, vinblastine, vincristine, etoposide, 5-FU, etc.)
Neutropenia (bone marrow damage)
Nose bleeds (isotretinoin)


Optic atrophy (vincristine)
Ototoxicity (cisplatin)


Photophobia (pentostatin)
Paralytic ileus (vinblastine)
Peripheral neuropathy (vinorelbine)
Phlebitis (BCNU, docetaxel)
Proteinuria (suramin)
Pulmonary fibrosis and toxicity and respiratory distress (esp. bleomycin, busulfan, carmustine, cyclophosphamide)


Quality of life, diminished or destroyed


Red urine (daunorubicin, doxorubicin) or orange urine (amonafide)


Seizures (BCNU)
Skin desquamation (paclitaxel)
Sterility and reproductive system dysfunction such as menstrusal and spermatogenesis dysfunction (busulfan, chlorambucil, cyclophosphamide, vincristine)
Stomatitis (nitrogen mustard, chlorambucil, etc.)


Taste perversion (taxotere)
Thrombocytopenia (esp. mitomycin and nitrosureas)
Tumor lysis syndrome (fludarabine)


Ulcerations of GI tract (5-FU)
Urticaria (paclitaxel)


Visual disturbance (mitotane)
Vomiting (severe with nitrogen mustard)
Venoocclusive disease (high-dose busulfan)


Weakness (vinorelbine)
Weight loss
Wheezing (leucovorin)


Xerostomia (isotretinoin)
X-ray or radiation recall: dermatitis in previously irradiated areas (dactinomycin)


Yeast infections


Zoladex-induced testicular atrophy
Zubrod Performance Status Scale decline

Finally, the following news item speaks for itself. It comes from the New York Times website (10/12/00):

"Genentech, Inc. on Wednesday reported a 27 percent gain in third-quarter profitsdriven by sales of cancer drugs, Herceptin and Rituxan. Third-quarter sales of Herceptin increased 52 percent, to $72.6 million Sales of Rituxan increased 62 percent, to $117.9 million."

In other words, just this one pharmaceutical company is now making over US $750,000,000 from the sale of two of its anticancer agents. The prediction I made at the German Society of Oncology meeting in 1997 that the cancer chemotherapy market would reach $13.8 billion by the end of the millennium now seems like a serious underestimation.

--Ralph W. Moss, Ph.D.

Director, The Moss Reports

 Generic  Brand name  Company  Indication  Approval Date
 anastrozole Arimidex Zeneca breast IV 1996
 docetaxel Taxotere Aventis breast IV, lung IV 1996
 mitoxantrone Novantrone Immunex prostate pain 1996
 irinotecan Camptosar Pharmacia colon IV 1996
 gemcitabine Gemzar Lilly NSCLC 1996+
 topotecan Hycamptin Smith-Kline ovarian 1996
 letrozole Femara Novartis breast IV 1997
 rituximab Rituxan Genentech lymphoma 1997
 porfimer sodium Photofrin QLT lung, esophagus 1998
 capectiabine Xeloda Hoffman-LaR breast 1998++
 temozolomide Temodar Schering brain (AA) 1999