Cancer in nonagenarians: Profile, treatments and outcomes☆

1. Introduction 

The world population is aging. In the US, people aged 85 and older are the fastest growing segment of the population: their number is expected to quadruple over the next 40–50 years. By year 2050 over 20 million of Americans are expected to be in that age range (http://www.census.gov/compendia/statab/tables/08s0010.xls) (accessed 6/26/08). The incidence of cancer increases steadily with age, and therefore a high prevalence of cancer can be expected in this population. The prevalence may plateau at the oldest ages. In the New England Centenarian study, the prevalence of cancer is 22%, whereas the lifetime risk would be expected to be 34% based on the SEER data [1]. So longevity phenotype may be associated with a lower incidence of cancer. However, this still remains a significant number. Very little information is available to guide the oncologists in the treatment of the oldest old. This seems to result in strikingly low rates of treatment in Veteran patients aged 85 and above [2]. Among patients with lung, head and neck, colorectal and prostate cancer, only 0.6% underwent surgery, 0.7% radiation therapy and 0.8% chemotherapy in 2005. In a population database, breast cancer patients above the age of 80 were often undertreated, with a negative impact on cancer-specific survival [3]. Patients aged 85 and above were notably less likely to receive surgery or radiation therapy. Women >85 were more likely to have poor health status and comorbid conditions. However only 17% of the women who did not undergo surgery were in poor health. Similar results were found in hospital databases [4]. Although several series reviewed the treatment of patients in their eighties, these reports contain very few patients in their nineties (e.g., [5], [6], [7], [8], [9], [10]). Most also focus on a single modality or disease. The most frequent conclusion is that selected patients can undergo safely various cancer treatments, but the numbers of eligible patients decrease markedly as age advances. In order to provide broader clinical data about effectiveness and tolerance to treatment in nonagenarians, we decided to review the experience of a large academic cancer center in the US: H. Lee Moffitt Cancer Center.

2. Methods 

Using Moffitt's cancer registry, we reviewed the charts of patients registered at Moffitt Cancer Center who were aged 90 or older at the time of treatment/evaluation. We retrieved data on cancer diagnosis and the treatments received at Moffitt past the age of 90. Treatment past age of 90 meant either newly diagnosed cancers, or the first line of treatment received at Moffitt for a cancer previously treated outside. The period covered was 7/1/1993–3/1/2006. Patients were included if they had a diagnosis of cancer, a clear treatment plan with two follow-up visits or more over a 1 month time period at Moffitt, or 1 evaluation with a treatment plan and death occurring within 6 months of their evaluation. Patients in this age range who were in follow-up for cancer treated at a younger age at Moffitt were not included. Data on the patient's age, gender, comorbidities, cancer histology and stage, treatments, and outcomes were extracted. Date of death was extracted from the Moffitt Cancer registry, supplemented with social security public data. In this descriptive cohort study, descriptive statistics were used as appropriate, and survival analysis was performed using the Kaplan–Meier method. Confidence intervals were reported for the median survival and the survival estimates. This study was approved by the Institutional Review Board of the University of South Florida.

3. Results 

Out of 580 patients aged 90 and older visiting Moffitt during this period, 177 were eligible for inclusion in this study. The primary reasons for exclusion were lack of cancer diagnosis (more than half of the patients were simply referrals to our speech therapy program for swallowing studies), lack of follow-up (2nd opinions or consultations), and no active treatment (patients treated at Moffitt at a younger age simply coming for follow-ups). The patient characteristics and types of cancer are listed in Table 1. It is interesting to note that 23.5% of patients had a previous or concomitant diagnosis of a second cancer. Our data did not allow to reliably assess the mean interval between the two cancers. Six patients (3.4%) had more than two cancers. The patients had for the most part a good performance status (0 or 1), as estimated by the ECOG PS. Although this might represent a referral bias, it should also be noted that ECOG PS estimates have a poor sensitivity to functional impairment in the elderly [11]. The number of concomitant medications was comparable to that of the Senior Adult Oncology Program patients during this period [12], and the use of complementary medicines was low: less than 10%. There is a time trend with the completeness of comorbidity reporting: the older charts often provide few details. With this taken into account, 31% of patients had no/unknown comorbidities, 14% 1 comorbidity, 17% 2 comorbidities, 19% 3 comorbidities, 12% 4 comorbidities, 3% 5 comorbidities, 3% 6 comorbidities, and 1% 7 comorbidities. Among the 136 patients with a comorbidity reporting deemed reliable, the median number of comorbidities was 3, which corresponds to that found in the general Senior Adult Program population [11]. The most frequent comorbidities were cardiovascular diseases. Forty-three patients had had treatment for their current cancer prior to presenting at Moffitt.

Table 1. Patient profile (177 patients).

	Age: median (range)	91.5 (90–102.66)
	Sex: M/F	66 (37.1%)/112 (62.9%)
	ECOG PS: 0/1	168 (95%)
	≥2	9 (5%)
	aComorbidities: median (range)	3 (0–7)
	aPrevious cancer	32 (23.5%)
	Colon	8
	Prostate	6
	Contralat. breast	3
	aMedications: median (range)	4 (0–14)
	aComplementary medicines	13 pts (9.6%)
	Cancer type
	Breast	25 (14.1%)
	Melanoma	24 (13.6%)
	Skin, otherb	22 (12.5%)
	Head and neck	15 (8.5%)
	Colorectal	15 (8.5%)
	Prostate	12 (6.8%)
	Gynecologic	11 (6.2%)
	Lung	10 (5.6%)
	NHL	10 (5.6%)
	Sarcoma	8 (4.5%)
	Hematologic, other	7 (4%)
	GU, other	6 (3.4%)
	CUO	5 (2.8%)
	GI, other	5 (2.8%)
	CNS	2 (1.1%)
	Cancer stage
	Solid tumors: I/II	46 (41%)
	III/IV	65 (59%)

a 136 patients evaluable. b Only invasive/metastatic cancers requiring surgeon/radiation/systemic treatment were included.

The treatments received are detailed in Table 2. Surgery comprised 37% of the treatments, 16% were radiation therapy, 12% chemotherapy, 11% hormonal therapy, and 7% targeted therapies. Our cohort comported one centenarian, who received palliative radiation for a head and neck cancer. As a first line treatment, the most frequent treatment was surgery (73 patients, 14 as second line, and 3 as third line), with 17 patients undergoing major surgery (abdominal or thoracic), and 12 other receiving head and neck surgery. Surgery overall had a low mortality rate, with only one patient dying from post-operative flutter and hypotension. Two additional patients had a post-operative myocardial infarction but recovered. One patient had to be reintubated because of hypotension but also recovered. One head and neck cancer patient developed an osteonecrosis of the jaw later in the course of disease. One patient could not receive adjuvant treatment because of poor post-operative performance status.

Table 2. Treatments received.

	Treatment received	Prior outside treatment (for same cancer)	1st treatment @ Moffitt	2nd	3rd and more
	Surgery	21	73	14	3
	Major (abd or thoracic)	–	17
	Chemotherapy	5	16	7	7
	Radiation	5	27	9	3
	Concurrent Chemo/Radiation	2	4	2	–
	Hormonal manipulation	7	21	4	2
	Hospice	–	14	3	1
	Clinical trial	–	1	–	–
	Targeted therapy	–	7	3	1
	Obs/supportive/recommendation	–	15	1	–
	Other	3	1	1	1

Thirty-nine patients did receive radiation therapy, which was overall well tolerated.

Thirty chemotherapy treatments were given, 16 as a first treatment at Moffitt, and 7 each as a second or third treatment. Most treatments were single agent chemotherapies or weekly regimens, but more intensive regimens such as VNCOP-B, Carboplatin/paclitaxel q3w or CHOP were given (one patient each). About half of the patients were treated with standard doses and half with reduced dose regimens. As major complications of chemotherapy, one patient died of pneumonia during her third line of chemotherapy for CLL. Three other patients developed severe infections. One patient was hospitalized for congestive heart failure on trastuzumab and vinorelbine, but recovered rapidly, and underwent successfully two lines of single agent chemotherapies. One patient developed a poor performance status on gemcitabine given immediately after outside chemo-radiation therapy for lung cancer. One patient received a 7+3 regimen for AML prior to coming to Moffitt, went into complete remission, but developed severe deconditioning preventing consolidation. She recovered her strength 4 months later. She relapsed one and a half year later and received low-dose Ara-C. Another AML patient, with M3 disease, was treated with ATRA alone, went into remission for 20 months, and then was treated with arsenic trioxide for relapse.

Four patients received concomitant chemo-radiation therapy as their initial Moffitt treatment, and two received it as a second line treatment. A 95 years old man with localized stage IV squamous cell cancer of the tongue received radiation with weekly cisplatin 20mg/m2. Cisplatin was stopped after 2 doses and he could complete radiation with excellent tumor response. Another 93 years old man with a large skin squamous cell cancer and lymph node invasion received radiation and weekly cisplatin, tolerated it well and survived at least 15 months. A 92 years old woman with recurrent squamous cell cancer of the lip was treated with radiation and concurrent weekly carboplatin, AUC of 1, with an excellent response that lasted 14 months. A 90 years old man with rectal cancer was enrolled in a study of concomitant oral topotecan and radiation, underwent low anterior resection, and was still alive 33 months later. A fifth patient, a man with locally advanced rectal cancer, received initially radiation therapy alone at the age of 90. A year later he relapsed and was treated with concomitant radiation and 5-FU/leucovorin. He had an excellent response for 3.5 years, then relapsed and was referred to hospice. A 90 years old man underwent transurethral resection and chemotherapy for bladder cancer, relapsed, and after failure of two resection attempts, received radiation with weekly carboplatin (AUC 2) and paclitaxel (30mg/m2). He developed renal failure under treatment and was referred to hospice.

The median survival of our patients was 1.69 years (95% CI=1.34–2.17, range 0.01–6.21 years), with 23% of them still alive at 4 years (Fig. 1). The numbers are barely changed if one excludes the patients with advanced skin cancers. Among the patients with solid tumors, the median survival of our cohort was 2.02 years [95% CI=1.56, 2.87] for patients with stages I–II cancers, and 1.06 years [95% CI=0.58, 1.63] for patients with stages III–IV cancers (p=0.02 by log-rank test) (Fig. 2).

View Large Image Download to PowerPoint

Fig. 1. Overall survival.

View Large Image Download to PowerPoint

Fig. 2. Median survival by stage (solid tumors).

4. Discussion 

The number of nonagenarians or even centenarians seeking treatment for cancer is increasing. Yet very few data are available to guide our treatment. Our study constitutes the largest report of a nonagenarian treatment cohort to date. Several studies exist that report outcomes of patients aged 80 and older, but these series contain only a small minority of nonagenarians, if any (e.g., [3], [4], [5], [6], [7], [8], [9], [10]). We were unable to find in the literature a clinical series focusing exclusively on cancer patients aged 90 years and older. One epidemiological series contain some comments on treatment [13].

Our study highlights the following points:

Nonagenarians are mostly referred for surgery and radiation therapy. Fifteen percent of our patients received some form of chemotherapy. Only one patient was enrolled in a clinical trial. Another was screened but in the end unable to be enrolled. This may be a combination of a large number of patients receiving surgery or radiation therapy alone and serious restrictions placed on nonagenarian patients by the studies’ inclusion criteria.

Strikingly, 23.5% of our patients had a second tumor. In the vast majority of cases, it was a preceding tumor, and two patients had concomitant cancers of the skin. Seven of these previous cancers were colorectal cancers. This is a higher proportion of colorectal cancers than what is encountered in our younger cohorts in their seventies or eighties and would deserve further exploration [14]. This high prevalence of second tumors has clear implications for the follow-up of older cancer survivors. Active screening and diagnostic work-up for a second cancer should be implemented, and new disease simply cannot be assumed to be a metastasis of a previous cancer.

Our study demonstrates nevertheless that nonagenarians are a frail population. The median survival of a 90 year old subject is 4.8 years in the US, and 3.3 years for a 95 years old subject (http://www.census.gov/compendia/statab/tables/08s0099.pdf, accessed 7/23/08). In our population, it was 1.69 years. It is important to note, however, that even in this population with limited survival, advanced stage cancer has a life-limiting effect (Fig. 2). It is also noteworthy that a quarter of our patients were still alive at 4 years. The 5-year survival of nonagenarians in the SEER registry is 12%, a figure within the range of our own numbers (Fig. 2) [13]. A review of the SEER data found that, after the first year, the annual relative survival of nonagenarians with cancer was not different than that of younger patients, up to 10 years of follow-up [13]. The excess mortality in the oldest old was during the first year, especially the first 4 months. This seems indeed compatible with the fact that in our series, there is a 24% 6-month mortality. Therefore, there appears to be two significant subgroups of patients: those who will die rapidly, either from an aggressive or advanced cancer, or because cancer is the decompensating event in a frail patient, and another subgroup who can achieve a few years of survival. Our data, contrary to those from the SEER registry, do not support the idea that treatment related mortality is an important cause of early death in nonagenarians. On this aspect, the fact that this series represent patients treated in a NIH-designated Comprehensive Cancer Center might be significant. Another significant aspect is that the SEER data covered a period from 1973 to 1998, compared to 1993 to 2006 in our series, and that anesthesia, surgery, radiation, and chemotherapy techniques have all made significant progress during that interval. Another reason for our good outcomes might be that, as we are a specialized center without an emergency room, most of our surgeries are elective. In nonagenarians undergoing general surgery, there is a large difference in 30-day mortality between emergency surgery (45%), urgent surgery (16%), and elective surgery (2.3%) [15]. Although the latter series dates from the 1980s, emergency surgery still remains a risky proposition in the very old [16]. Despite scarce data in nonagenarians, other series confirm that radiation therapy is well tolerated in the elderly [5].

Our study has some limitations, in addition to those mentioned above. As a tertiary cancer center, we likely have a selected referral population, with the sickest nonagenarians being treated locally. However, the performance status and comorbidity level of our patients is comparable to that of our SAOP patients in their seventies and eighties. Once patients are referred to us, their treatment tolerance seems to be good. In this study, the survival times are calculated from the time of first consultation at Moffitt, which typically occurs after diagnosis. This would tend to give a lower estimate of survival than the survival from diagnosis. About a quarter (24%) of the patients had prior treatment for their present cancer outside Moffitt. Thus, in most cases, our patients were seen early in their treatment.

In conclusion, selected nonagenarians with a good performance status can receive the full array of oncologic treatments. Their median survival, as a group, is a couple of years, and therefore the major emphasis should be on early benefit from the treatment. A significant minority of individuals will derive long-term benefit (4 years and more) from their treatment.

Conflict of interest statement 

The authors have no conflict of interest to disclose for this manuscript.

Reviewers 

Supriya Gupta Mohile, M.D., M.S., Assistant Professor, University of Rochester, Department of Medicine, 601 Elmwood Avenue, Box 704, Rochester, NY 14642, United States.

Cynthia Owusu, M.D., Case Western Reserve University, Department of Hematology/Oncology, 11100 Euclid Avenue, Cleveland, OH 44106, United States.