Origins of timed cancer treatment: early marker rhythm-guided individualized chronochemotherapy

Abstract

A 21-year old patient who presented in 1973 with a rare and highly malignant ovarian endodermal sinus tumor with spillage into the peritoneal cavity is alive and well today after receiving chronochemotherapy. During the first four courses of treatment, medications were given at different circadian stages. Complete blood counts and marker variables such as mood, vigor, nausea, and temperature were monitored around the clock and analyzed by cosinor to seek times of highest tolerance. Remaining treatment courses were administered at a time corresponding to the patient's best drug tolerance, rather than extrapolating the timing of optimal cyclophosphamide administration from also-implemented parallel laboratory studies on mice. Notwithstanding remaining hurdles in bringing chronochemotherapy to the clinic for routine care, merits of marker rhythm-guided chronotherapy documented in this and other case reports have led to the doubling of the two-year disease-free survival of patients with large perioral tumors in a clinical trial.

AIM

To document the value of timing the treatment of cancer by presenting a patient (CN) who in the 1970s underwent surgical removal of a rare and highly malignant ovarian endodermal sinus tumor with spillage into the peritoneal cavity, and who is alive and well more than 30 years after subsequent chronochemotherapy. This case involved instructive temperature, mood, vigor and hematological marker-guided rhythmometry, prompting a systematic trial of timed radiotherapy, which doubled the 2-year disease-free survival rate of patients with perioral cancers.

BACKGROUND OF TIMED TREATMENT

CN had a very poor prognosis at the time of admission, with a 10% 2-year survival rate (1). This report follows earlier autorhythmometry-guided timed treatment (2, 3). It precedes the first prospective randomized trial of tumor temperature-guided chronoradiotherapy which doubled the 2-year disease-free survival of patients with very advanced perioral cancers. Statistically validated results were demonstrated by cosinor (4) and subsequently illustrated as tumor regression in the time domain (5; see Figure 2 bottom middle therein), Figure 1. Other endeavors toward individualizing tumor marker rhythm-guided initiatives for the timed treatment of cancer followed, Figure 2. Disease recurrence in the setting of an increase in a cancer marker within the inappropriately-named “normal” range that was not aggressively treated has been studied in great detail and has led to designs for the individualized optimization of cancer treatment timing (6).

Figure 2.

Extent of within-day change and extent of its reproducibility in a given patient receiving treatment timed by marker rhythmometry, which added years of life to a case described by her oncologist as benefiting from ”chronopharmacologic principles [that] resulted in a dramatic objective remission of the disease. A three-year survival is far beyond the average period of control for advanced ovarian cancer.” (Kennedy BJ. A lady and chronobiology. Chronobiologia 1993; 20: 139−144.)

Figure 1.

Clinical outcomes in lower middle section show merit of radiation treatment at circadian peak tumor temperature.

CASE REPORT

CN, a G₀P₀ 24-year-old Caucasian woman, was first noted to have a left pelvic mass by palpation on February 7, 1973, at the age of 21 years, during a follow-up examination for a Copper T intrauterine contraceptive device which had been inserted two months previously. Prior to the follow-up examination, the patient had noted an increase in abdominal girth for the previous half-month and crampy and sharp left lower quadrant pain for the preceding week. On February 16, 1973, CN was admitted to Emmanuel Hospital, Portland, Oregon, where an exploratory laparotomy revealed a right ovarian tumor, with both the right ovary and tube pushed to the left side.

A right salpingo-oophorectomy, left ovarian bivalve procedure, and an incidental appendectomy were performed. During the procedure, the cystic tumor contents were inadvertently spilled into the peritoneum. The pathologic report revealed an 8-cm endodermal sinus tumor (yolk sac carcinoma) of the right ovary. CN recovered from surgery without complications and was referred to the University of Minnesota for chemotherapy.

CN was first admitted to the University of Minnesota on March 6, 1973, and on March 13, 1973, was started on combination chemotherapy consisting of cyclophosphamide (Cytoxan) 7 mg/kg IV qd × 5 days a month for 20 months, actinomycin D 0.5 mg qd × 5 days q month for 20 months, and vincristine 1.5 mg/m² IV q week × 12, the latter, however, being discontinued after 3 doses because of abdominal pain and paresthesis attributable to vincristine toxicity. The timing of drug administration was varied from month to month for the first 4 months and autorhythmometry (mood, vigor, nausea, temperature) and complete blood counts were followed to determine the patient's time of highest tolerance. Prior to each course of medication, blood samples were drawn q4 h for 24 hours at specific timepoints to determine the circadian rhythm of circulating platelets, WBC and differential, hemoglobin, hematocrit and RBC indices. Core temperature, blood pressure, pulse, mood, and vigor were also assessed q4 h for the next 24 hours. Then, at scheduled times subsequent to medication, WBC were obtained to assess bone marrow suppression. Oral temperature was measured 5 or more times a day, on most days, mostly during wakefulness, Figure 3. Regularly, around the clock, mood and vigor were self-rated by admittedly subjective criteria. In March, the medications were given between 0800−1100, in April between 2000 and 2200, in May at 0400 and in June again at 2200. After the fourth course, 0400 was chosen as the preferred time of administration based on criteria for the patient's drug tolerance, rather than by extrapolation of optimal cyclophosphamide timing from mice (and then correcting for differences in diurnality vs. nocturnality of activity) (7), Figure 4. CN received 20 courses of treatment covering 19 months..

Figure 3.

Illustrative, albeit bleached record of the computer display summarizing self-measurements of vigor (top) and oral temperature (below). In each of the two sections, probabilities are given on top for the test of the zero-amplitude (no-rhythm) assumption, with P-values below 0.05 seen below the dashed horizontal line: relatively few values are statistically significant for vigor; more values are below the 5% line for temperature in the upper section of the bottom half. Of the two curves in each section, the lower curve is the MESOR and the distance between the two curves is the amplitude, which is very small at the outset but gets measurably bigger for oral temperature and is sometimes statistically significant toward the end of the treatments. Graphs reveal the extent of disciplined self-measurements and inferential statistical computer analyses implemented in the early 1970s.

Figure 4.

Circadian susceptibility rhythm to cyclophosphamide in mice (pooled data from 5 separate studies on a total of 706 mice analyzed by cosinor). CL: confidence limits. From (7).

COURSE DURING 19 MONTHS OF CHRONOCHEMOTHERAPY

The patient developed alopecia and psoriasis, in addition to nausea and vomiting with each course of therapy, necessitating the use of IV fluids to maintain hydration. Prophylactic Ampicillin was used briefly after the second course of therapy because of leukopenia (WBC 1650). The WBC was never found to go below 1100.

Her course was complicated by an acute bronchitis in January 1974, treated with erythromycin, although no pathological microorganism was identified, and keratoconjunctivitis in October 1974, which resolved with sulfacetamide and maxidex. In December 1973, the patient fractured her left radial head in a fall, and in March 1974, a benign right T10 anterior rib fracture was diagnosed. The patient was rendered post-menopausal.

In December 1974, about 5 weeks after the 20^th course of chronochemotherapy, the patient underwent a second look operation at the University of Minnesota, and no evidence of tumor recurrence was found. Peritoneal washings were negative. Adhesions between ileum, colon and the site of the previous salpingectomy were removed and revealed no evidence of tumor. The left ovary was noted to be small with no evidence of recent follicular activity, supporting the diagnosis of premature menopause. The patient recovered from surgery without complication.

COURSE POST-THERAPY

Since December 1974, the patient has been followed at the University of Minnesota Gynecologic Oncology Clinic. She has had no evidence of recurrence. Her AFP, CBC and chest X-ray have been followed and remain within acceptable limits. She functioned well again in her original occupation, and later became a nurse.

In January 1997, the patient was treated for a major depression related to recent losses of cancer patients she had cared for, to erosion of standards in nursing care, the struggle to balance competing demands, and recalling her own cancer. Her depression resolved, and as of 2006 she is in good health.

In a recent telephone follow-up, the patient welcomed this write-up of her case. While today yolk sac tumors are generally responsive to currently available chemotherapy (8), her case at the time had a very poor prognosis with treatments then available (1). “Timing” her therapy was less based on the best time extrapolated from concomitant studies on laboratory animals (7), Figure 4, and mostly on actually changing the time of chemotherapy based on objective comparisons of variables related to hematotoxicity and general well-being, complemented by around-the-clock mood and vigor self-rating.

A second case of successful chronochemotherapy involved a patient who was convinced of the merit of timing treatment. He had read the literature and could interpret the successes in laboratory animals (9-11), understanding that a contrary view (12) stemmed from an unwarranted interpretation of the data, with all cured animals being omitted from the comparison of survival times (12). The data convinced him to perform marker rhythmometry on himself, and he changed his schedule of living to be able to get chemotherapy at a time convenient to the oncologist, since chronobiologically-timed treatment could not be scheduled. By changing his sleep-wakefulness and meal schedules, he shifted his circadian rhythms so that chemotherapy could be administered at a time corresponding to the putative optimal time gauged by his body temperature rhythm. Today, cancer markers are known to undergo multifrequency rhythms, some of which have been mapped, e.g., Figure 2 (cf. also Figure 3 in 13; cf. 14). It seems reasonable to advocate their use, preferably up front (5), notably when clock-hour-based timed treatment recommendations, that can be successful (15), have failed (16, 17; cf. 18).

Cost could almost certainly be reduced if the merit of using marker rhythms, based on around-the-clock longitudinal sampling (6, 19) for timing optimization and validation of effect would be more generally recognized. B.J. Kennedy, the founder of the specialty of oncology in the USA, estimated that chronobiology (with very intensive cancer marker rhythmometry), Figure 2, added several years of life in another case with an extremely poor prognosis (20; cf. 5, 19). While repeated blood sampling is obviously not recommended in cancer patients who may already be anemic, repeated sampling of saliva and/or urine is easier and can be useful in guiding treatment timing (5, 6, 19, 20).

Against this background, the continuous surveillance and analysis of temperature remains a more affordable and more practical approach but only as a complement, rather than as an alternative for tumor markers, for which there is no substitute for diagnosis. In validating the effect of therapy, thermometry is particularly attractive since there is evidence that measuring core temperature may approximate tumor temperature (21, 22), at least when it is used as a guide for treatment timing, even though its usefulness as a gauge of treatment efficacy has not been validated. One remaining question deserving further study, however, is that of the desirability of adjusting treatment time for repeated courses of treatment for a comparison of effects, notably when studies have shown that the circadian temperature rhythm may become desynchronized from 24 hours in the presence of cancer (23-25). Chronotherapeutic trials (26; cf. 27) and laboratory studies (28) suggest that alternating treatment times during consecutive courses of chemotherapy may be beneficial, and it might have contributed to the cure of the patient discussed herein.

HURDLES EXIST FOR FURTHER RESEARCH

At a meeting of the Gynecologic Oncology Group in Minneapolis, Prof. Leo Twiggs, then head of the Obstetrics and Gynecology Department at the University of Minnesota, saw to it that one of us (GC) could introduce a proposal for marker rhythm-guided chronochemotherapy for ovarian cancer. By then, our proposal for treating ovarian cancers could be guided by tumor markers, CA130 vs. CA125 (14), Figure 2 . The response of oncologists was positive, but in a private consultation, the NIH statistician volunteered that the implementation of the proposal was not possible because he would have to add another page of inquiry to paperwork which already exceeded his capabilities!

Other relatively insignificant obstacles have also stymied the progress of chronobiology. In another example, when the Daiichi pharmaceutical company of Japan generously provided us with a gift of $50,000 worth of CA130 kits, the gynecologic oncology fellows in the University hospital were enthusiastic. Erhard Haus, head of laboratories at St. Paul-Ramsey Medical Center, was in a position to help and was most cooperative in carrying out the CA130 determinations, as he did in another case (6). Interest in the project notwithstanding, it was not implemented since transportation of the samples from the University hospital to Dr. Haus’ laboratory was the major hurdle.

We do not know whether the patient we report is alive because of chronochemotherapy. A single case report is necessarily anecdotal. A well-conducted chronoradiotherapy study, however, subsequently documented the doubling of 2-year survival (4, 5). The still broader evidence from laboratory animals (e.g., 9−11), clearly demonstrates that cancer chronotherapy can statistically significantly improve outcome, and is an important challenge to both care recipient and caregiver in order to improve cure rates. Marker rhythmometry can now guide the treatment of prehypertension such as circadian overswinging and that of MESOR-hypertension (29-32). In dealing with cancer or with the circulation, we must not “fly blind” (29).