Special Feature

Secondary Malignancies Induced by Radiation Exposure: Clinical Examples and Current Experimental Findings

By Roy Vongtama and David A. Corral, MD

Radiation therapy is a proven, effective modality for the treatment of many forms of cancer. Since its introduction as part of the standard of oncologic care, radiation has also been questionably associated with the evolution of secondary malignancies. Secondary malignancies are a serious and potentially lethal concern not only for the patient, but also for the physician, who must take this issue into account when deciding whether to recommend radiation as a part of therapy. The process by which radiation induces neoplastic transformation is not well understood. However, significant advances have been made in recent years in the basic research of radiation-induced malignant transformations which have indicated possible underlying mechanisms. This article will briefly explore the incidence of secondary malignancies in several organ systems and examine the current experimental theories of the mechanism of radiation-induced malignancy.


Radiation has been suggested as a potential cause of nearly every type of malignancy. With this being said, however, two facts must be recognized that significantly modify this statement. The first is that the levels of radiation exposure that have been shown to be clearly associated with secondary malignancies range from one to 10 Gray (Gy), exposure that is much lower than dosages typically used for therapy. Within this range of radiation dosages, a linear dose response between exposure and cancer incidence can be seen. Studies have failed to show the same linear dose response with the higher (therapeutic) levels of radiation, although increased risk has been demonstrated.2 It has been postulated that at therapeutic dosages, the high level of cell death may result in the destruction of potentially malignant cells that would otherwise have been transformed by the radiation. The second fact which must be recognized is that certain malignancies seem to be more radiosensitive than others. This article focuses on thyroid cancer, breast cancer, and soft tissue sarcomas—three malignancies that are highly radiosensitive.

Thyroid Carcinoma

Thyroid carcinoma is a classic example of a malignancy that has been shown in numerous studies to be induced by radiation exposure. In a study by Schneider and colleagues, 5.9% of patients who had received radiation therapy for benign head and neck conditions developed a secondary thyroid cancer 3-42 years after radiation therapy, with a median time to diagnosis of 10 years following radiation. Interestingly, the clinical course of the radiation-associated thyroid carcinomas did not vary from thyroid cancer diagnosed in patients with no prior radiation exposure. In children, the risk of radiation-induced thyroid cancer inversely correlated with their age at the time of exposure. Those who were exposed to radiation before age 5 had a five-fold higher incidence of secondary thyroid cancer than those exposed after 10 years of age.4 As seen with adults, the clinical course did not vary significantly from that of non-radiation-induced thyroid carcinomas, with the majority of diagnoses peaking 15-20 years after radiation.5

Breast Cancer

Radiation has also been suspected as a cause of breast cancer. Similar to thyroid carcinoma, the risk appears to be inversely proportional to age, with children exposed to radiation at a young age having a higher risk later in life. Within the low-dose range, it has also been shown that there is a linear relationship between dose and relative risk. Again, the clinical course of these radiation- associated breast cancers seems to be no different than that seen in breast cancer patients who had not been previously exposed to radiation.6

Soft Tissue and Bone Sarcomas

Sarcomas of soft tissue and bone associated with radiation exposure are well reported in the literature.7,8 For example, Souba and colleagues reported 16 cases of sarcoma arising in previously irradiated sites located in the chest wall. Because these patients received radiation therapy for nonsarcomatous tumors, these data support the theory that radiation can play a significant role in the oncogenic process of secondary malignancies. In general, as with non-radiation-associated sarcomas, radiation- associated sarcomas have a poor prognosis due to their normally advanced stage upon discovery, with the only effective therapy being surgical removal, and average five-year survival reaching only 30%.7

Experimental Model for the Study of Radiation-Induced Tumors

As can be seen from the above examples, a case can be made that radiation is a potent inducer of cancer, especially at low levels. However, this has yet to be proven convincingly in preclinical studies. Much of the current knowledge behind radiation-induced malignant transformation has been discovered in the last 10 years. The model discussed in this article was developed by Mendonca and associates and used a hybrid line of cells consisting of a tumorigenic cell line (HeLa) combined with human skin fibroblasts. This model is an excellent experimental model for the study of radiation-induced tumors for two reasons. As discussed above, many radiation-associated malignancies are sarcomas—tumors that result from malignant transformation of mesenchymal cells such as fibroblasts. Also, the model follows Knudson’s "two hit" hypothesis of carcinogenesis that was initially described in the retinoblastoma model. Briefly, Knudson’s hypothesis states that a cell with a genetic locus (encoding a tumor suppressor) which contains a mutated allele may be phenotypically normal as long as a nonmutated allele is normally expressed. For malignant transformation to occur, the second allele must be altered. This results in a loss of heterozygosity and allows the pre-existing genetic mutation to lead to malignant transformation.

The HeLa line used for the hybrid cell line has been shown to be inherently tumorigenic and produces the antigen intestinal alkaline phosphatase (IAP), which can be used as a marker to signal tumorigenic activity. The human skin fibroblast line has the characteristics of being nontumorigenic and having tumor-suppressing activity. Combined together, the hybrid line, HeLa* human skin fibroblast, is an actively tumor suppressing cell line that contains on average four alleles per chromosome, two from the nontumorigenic fibroblast and two from the tumorigenic HeLa line. The production of IAP, the HeLa neoplastic marker, is actively suppressed by gene products of the normal human skin fibroblast. This system provides a precise means to detect neoplastic transformation in the hybrid cell line (i.e., if the fibroblast’s tumor suppressor alleles are knocked out, the production of IAP will resume and can be quantitated).9,10 (Note: The actual hybrid cell line used for the experiment, CGL1, is a descendant of the original hybrid HeLa* fibroblast line.)

In the Mendonca et al study, hybrid cell lines were irradiated with 7 Gy of gamma rays at 2.1 Gy/min and then trypsinized, counted, and seeded into flasks containing standard growth medium. These cells were then fed, plated, and left to grow for 7, 9, 11, 13, 15, 17, 19, and 21 days, at which time they were fixed and stained for IAP with Western Blue reagent. If a cell had been transformed by the radiation exposure, it would begin to produce IAP. Neoplastic frequencies for each day (7-21) were measured by counting the number of neoplastic foci per total number of surviving cells. A higher frequency translated into a higher rate of malignant transformation. Interestingly, IAP positive foci did not begin to appear at a significant frequency until the 11th day. The IAP positive finding demonstrated that malignant transformation did occur, but occurred in a delayed fashion on day 11. This coincided with the reduced plating efficiency of cells seen, which also began on the 11th day. Plating efficiency refers to the ability of transferred cells to grow in appropriate media. With regard to plating efficiency, nonirradiated controls plated out at a relatively constant rate of 60-80%. The irradiated samples initially start at 11% on day 4, recover to 35-45% on day 9, and steadily decline thereafter. The reduced plating efficiency of irradiated cells demonstrated that a significant proportion of cells were unable to successfully replicate, possibly implying that they had undergone significant genetic damage that manifested itself on day 11.

The temporal coincidence of IAP appearance (i.e., malignant transformation) and reduced plating efficiency (i.e., genetic damage) on day 11 was a significant discovery that led to additional investigations into the mechanism of each. It was discovered that the tumor suppressor loci that were lost with radiation exposure were found to be on fibroblast chromosomes 11 and 14.9. These two loci were found to be necessary but not sufficient for neoplastic transformation, as control lines that had lost either of these suppressor loci did not develop neoplastic characteristics (namely, IAP production). Mendonca et al proposed that the delayed loss of tumor suppressor function was not a discrete event, but rather the buildup of heritable damage over the course of replicative cycles in which the final result was loss of tumor suppressor function.

Recent Laboratory Discoveries

In the most recent study published in Cancer Research, Mendonca et al investigated the significant reduction in plating efficiency seen in the irradiated cell population.10 It was demonstrated that these cells undergo a delayed apoptosis which causes the drastic reduction in plating efficiency. Apoptosis is a genetically mediated form of cell death induced by various stimuli. Radiation is known to be a potent epigenetic apoptotic stimulus.11 The upregulation of apoptosis around day 11 was demonstrated by several methods previously shown to be indicative of apoptosis, some of which will be listed here. DAPI staining was done, which indicated abnormal morphology consistent with apoptosis. Endonuclease-mediated DNA strand breaks consistent with apoptosis were also seen by TUNEL assays. Western blot showed increased levels of p53 and Bax, proteins which have been shown to be proapoptotic. In short, all techniques used were in agreement, thereby lending credence to the theory that plating efficiency was decreased because of induction of apoptosis. Mendonca et al proposed that the actual trigger for apoptosis is the same for loss of tumor suppressor function: it is not a single event, but rather the buildup of heritable damage over the course of replicative cycles.11 Termed genetic instability, the damage becomes great enough to trigger apoptosis.


With all this being said, the question remains: how does radiation induce neoplasia in these hybrid cells? A summary of the work of Mendonca et al gives a possible answer. Radiation is given, producing immediate cell death and also sublethal genetic damage. The remaining HeLa* fibroblast cells begin to replicate. The significant changes begin approximately on day 10; at this time, the hybrid cells have undergone approximately 12 population doublings. A buildup of genetic mutations occurs over these replicative cycles. This genetic instability can have two outcomes, one of which is the induction of apoptosis (and, therefore, cell death) that occurs for the majority of the cells. The second outcome is that a small subset of these irradiated cells lose their tumor suppressor genes (11 and 14) but either evade apoptosis or have not had enough accumulated damage to trigger apoptosis. Thus, they become neoplastic.

Patients who have been recommended to receive radiation as a primary form of treatment are often intrinsically afraid or at least aware of the potential for long-range side effects. This hesitancy is well founded: secondary cancers associated with radiation exposure are body wide and have a poor prognosis, as shown above. Clearly the work done by Mendonca et al provides significant advances in the quest to understand radiation-induced cancers. The future lies in further experimentation to better understand the mechanisms involved using more complex biological models. The ultimate goal is to understand the complete process of radiation-induced malignant transformation that will translate into clinical knowledge and help physicians determine the most effective therapeutic strategies that confer the least possibility for secondary malignancies. (Roy Vongtama is a medical student, School of Medicine and Biomedical Sciences, University at Buffalo; Dr. Corral works for the Department of Urologic Oncology, Roswell Park Cancer Institute, Buffalo, New York.)


1. Thompson DE, et al. Radiation Res 1994;137:S17.

2. Kaldor JM, et al. N Engl J Med 1990;322:7.

3. Schneider AB, et al. Ann Intern Med 1986;105:3.

4. Tucker MA, et al. Can Res 1991;51:2885.

5. Ron E, et al. Radiation Res 1995;141:259.

6. Finnerty NA, et al. Arch Intern Med 1984;144:1217.

7. Pitcher ME, et al. Eur J Surg Oncol 1994;20:53.

8. Souba WW, et al. Cancer 1986;57:610.

9. Mendonca MS, et al. Radiation Res 1998;149:246.

10. Mendonca MS, et al. Can Res 1999;59:3972.

11. Mendonca MS, et al. Radiation Res 1993;134:209.

Tumor suppressor loci lost in irradiated HeLa* fibroblast hybrids were localized to fibroblast chromosomes:

a. 9 and 12.

b. 10 and 13.

c. 11 and 14.

d. 12 and X.