Theory and practice of predictive assays in radiation therapy

Plasma Cell-Free DNA as a Predictive Marker after Radiotherapy for Hepatocellular Carcinoma

PURPOSE

Cell-free DNA (cfDNA) is gaining attention as a novel biomarker for oncologic outcomes. We investigated the clinical significance of cfDNA in hepatocellular carcinoma (HCC) patients treated with radiotherapy (RT).

MATERIALS AND METHODS

Fifty-five patients with HCC who received RT were recruited from two prospective study cohorts: one cohort of 34 patients who underwent conventionally fractionated RT and a second of 21 patients treated with stereotactic body radiation therapy. cfDNA was extracted and quantified.

RESULTS

In total, 30% of the patients had multiple tumors, 77% had tumors >2 cm, and 32% had portal vein tumor thrombus. Optimal cut-off values for cfDNA levels (33.65 ng/mL and 37.25 ng/mL, before and after RT) were used to divide patients into low-DNA (LDNA) and high-DNA (HDNA) groups. The pre-RT HDNA group tended to have more advanced disease and larger tumors (p=0.049 and p=0.017, respectively). Tumor response, intrahepatic failure-free rates, and local control (LC) rates were significantly better in the post-RT LDNA group (p=0.017, p=0.035, and p=0.006, respectively).

CONCLUSION

Quantitative analysis of cfDNA was feasible in our cohorts. Post-RT cfDNA levels were negatively correlated with treatment outcomes, indicating the potential for the use of post-RT cfDNA levels as an early predictor of treatment responses and LC after RT for HCC patients.

Ex Vivo Apoptosis in CD8+ Lymphocytes Predicts Rectal Cancer Patient Outcome

Background. Apoptotic rates in peripheral blood lymphocytes can predict radiation induced normal tissue toxicity. We studied whether apoptosis in lymphocytes has a prognostic value for therapy outcome. Methods. Lymphocytes of 87 rectal cancer patients were ex vivo irradiated with 2 Gy, 8 Gy, or a combination of 2 Gy ionizing radiation and Oxaliplatin. Cells were stained with Annexin V and 7-Aminoactinomycin D and apoptotic and necrotic rates were analyzed by multicolor flow cytometry. Results. After treatment, apoptotic and necrotic rates in CD8+ cells are consistently higher than in CD4+ cells, with lower corresponding necrotic rates. Apoptotic and necrotic rates of CD4+ cells and CD8+ cells correlated well within the 2 Gy, 8 Gy, and 2 Gy and Oxaliplatin arrangements (p ≤ 0.009). High apoptotic CD8+ rates after 2 Gy, 8 Gy, and 2 Gy + Oxaliplatin treatment were prognostically favorable for metastasis-free survival (p = 0.009, p = 0.038, and p = 0.009) and disease-free survival (p = 0.013, p = 0.098, and p = 0.013). Conclusions. Ex vivo CD8+ apoptotic rates are able to predict the patient outcome in regard to metastasis-free or disease-free survival. Patients with higher CD8+ apoptotic rates in the peripheral blood have a more favorable prognosis. In addition to the prediction of late-toxicity by utilization of CD4+ apoptotic rates, the therapy outcome can be predicted by CD8+ apoptotic rates.

Comparative analysis of three functional predictive assays in lymphocytes of patients with breast and gynaecological cancer treated by radiotherapy

Purpose

There is a need to develop predictive tests that would allow identifying cancer patients with a high risk of developing side effects to radiotherapy. We compared the predictive value of three functional assays: the G₀ aberration assay, the G₂ aberration assay and the alkaline comet assay in lymphocytes of breast cancer and gynaecological cancer patients.

Material and methods

Peripheral blood was collected from 35 patients with breast cancer and 34 patients with gynaecological cancer before the onset of therapy. Chromosomal aberrations were scored in lymphocytes irradiated in the G₀ or G₂ phase of the cell cycle. DNA repair kinetics was performed with the alkaline comet assay following irradiation of unstimulated lymphocytes. The results were compared with the severity of early and late side effects to radiotherapy.

Results

No correlation was observed between the results of the assays and the severity of side effects. Moreover, each assay identified different patients as radiosensitive.

Conclusions

There is no simple correlation between the in vitro sensitivity of lymphocytes and the risk of developing early and late side effects.

Prediction of clinical toxicity in locally advanced head and neck cancer patients by radio-induced apoptosis in peripheral blood lymphocytes (PBLs)

Head and neck cancer is treated mainly by surgery and radiotherapy. Normal tissue toxicity due to x-ray exposure is a limiting factor for treatment success. Many efforts have been employed to develop predictive tests applied to clinical practice. Determination of lymphocyte radio-sensitivity by radio-induced apoptosis arises as a possible method to predict tissue toxicity due to radiotherapy. The aim of the present study was to analyze radio-induced apoptosis of peripheral blood lymphocytes in head and neck cancer patients and to explore their role in predicting radiation induced toxicity. Seventy nine consecutive patients suffering from head and neck cancer, diagnosed and treated in our institution, were included in the study. Toxicity was evaluated using the Radiation Therapy Oncology Group scale. Peripheral blood lymphocytes were isolated and irradiated at 0, 1, 2 and 8 Gy during 24 hours. Apoptosis was measured by flow cytometry using annexin V/propidium iodide. Lymphocytes were marked with CD45 APC-conjugated monoclonal antibody. Radiation-induced apoptosis increased in order to radiation dose and fitted to a semi logarithmic model defined by two constants: α and β. α, as the origin of the curve in the Y axis determining the percentage of spontaneous cell death, and β, as the slope of the curve determining the percentage of cell death induced at a determined radiation dose, were obtained. β value was statistically associated to normal tissue toxicity in terms of severe xerostomia, as higher levels of apoptosis were observed in patients with low toxicity (p = 0.035; Exp(B) 0.224, I.C.95% (0.060-0.904)). These data agree with our previous results and suggest that it is possible to estimate the radiosensitivity of peripheral blood lymphocytes from patients determining the radiation induced apoptosis with annexin V/propidium iodide staining. β values observed define an individual radiosensitivity profile that could predict late toxicity due to radiotherapy in locally advanced head and neck cancer patients. Anyhow, prospective studies with different cancer types and higher number of patients are needed to validate these results.

Prediction of clinical toxicity in localized cervical carcinoma by radio-induced apoptosis study in peripheral blood lymphocytes (PBLs)

Background

Cervical cancer is treated mainly by surgery and radiotherapy. Toxicity due to radiation is a limiting factor for treatment success. Determination of lymphocyte radiosensitivity by radio-induced apoptosis arises as a possible method for predictive test development. The aim of this study was to analyze radio-induced apoptosis of peripheral blood lymphocytes.

Methods

Ninety four consecutive patients suffering from cervical carcinoma, diagnosed and treated in our institution, and four healthy controls were included in the study. Toxicity was evaluated using the Lent-Soma scale. Peripheral blood lymphocytes were isolated and irradiated at 0, 1, 2 and 8 Gy during 24, 48 and 72 hours. Apoptosis was measured by flow cytometry using annexin V/propidium iodide to determine early and late apoptosis. Lymphocytes were marked with CD45 APC-conjugated monoclonal antibody.

Results

Radiation-induced apoptosis (RIA) increased with radiation dose and time of incubation. Data strongly fitted to a semi logarithmic model as follows: RIA = βln(Gy) + α. This mathematical model was defined by two constants: α, is the origin of the curve in the Y axis and determines the percentage of spontaneous cell death and β, is the slope of the curve and determines the percentage of cell death induced at a determined radiation dose (β = ΔRIA/Δln(Gy)). Higher β values (increased rate of RIA at given radiation doses) were observed in patients with low sexual toxicity (Exp(B) = 0.83, C.I. 95% (0.73-0.95), p = 0.007; Exp(B) = 0.88, C.I. 95% (0.82-0.94), p = 0.001; Exp(B) = 0.93, C.I. 95% (0.88-0.99), p = 0.026 for 24, 48 and 72 hours respectively). This relation was also found with rectal (Exp(B) = 0.89, C.I. 95% (0.81-0.98), p = 0.026; Exp(B) = 0.95, C.I. 95% (0.91-0.98), p = 0.013 for 48 and 72 hours respectively) and urinary (Exp(B) = 0.83, C.I. 95% (0.71-0.97), p = 0.021 for 24 hours) toxicity.

Conclusion

Radiation induced apoptosis at different time points and radiation doses fitted to a semi logarithmic model defined by a mathematical equation that gives an individual value of radiosensitivity and could predict late toxicity due to radiotherapy. Other prospective studies with higher number of patients are needed to validate these results.

DNA double-strand break repair defects in syndromes associated with acute radiation response: at least two different assays to predict intrinsic radiosensitivity?

PURPOSE

Human diseases associated with acute radiation responses are rare genetic disorders with common clinical and biological features including radiosensitivity, genomic instability, chromosomal aberrations, and frequently immunodeficiency. To determine what molecular assays are predictive of cellular radiosensitivity whatever the genes mutations, the existence of a quantitative correlation between cellular radiosensitivity and unrepaired DNA double-strand breaks (DSB) repair defects was examined in a collection of 40 human fibroblasts representing 8 different syndromes.

MATERIALS AND METHODS

A number of techniques such as pulsed-field gel electrophoresis, plasmid assay and immunofluorescence with antibodies against MRE11, MDC1, 53BP1 and phosphorylated forms of H2AX, DNA-PK were applied systematically.

RESULTS AND CONCLUSIONS

Survival fraction at 2 Gy was found to be inversely proportional to the amount of unrepaired DSB, whatever the genes mutations and the assay applied. However, no single assay discriminates the full range of human radiosensitivity. Particularly, nuclear foci formed by the phosphorylation of H2AX do not predict well moderate radiosensitivities. Our findings suggest the existence of an ATM-dependent interplay between the activation of DNA-PK and MRE11. A classification of diseases according their cellular radiosensitivity, their molecular response to radiation and the functional assays permitting their evaluation is proposed.

Assessing the likelihood of severe side effects in radiotherapy

Ionizing radiation is a powerful tool to treat cancer. The curing effect is mainly based on the efficiency of ionizing radiation to kill the cancer cells and it is believed that DNA double-strand breaks (DSBs) represent the most significant genetic lesion introduced by radiation that causes cell killing. One limitation in radiotherapy is the unavoidable damage delivered to the normal, noncancer cells that can give rise to side effects. The ultimate goal in treatment planning is to maximize cell killing in the tumor by minimizing damage induction in the normal tissue surrounding the tumor. The biological response to the induction of DSBs is largely affected by DSB repair processes and it has, therefore, been a long-standing goal to determine a patient's DSB repair capacity to "individualize" treatment planning. A recently developed DSB repair assay that allows the assessment of patients' repair capacity under in vivo conditions may provide a new approach to predict individuals' responses to radiotherapy and may be able to contribute to improvements in treatment planning.

Sources of variation in patient response to radiation treatment

PURPOSE

To investigate sources of variation in radiosensitivity displayed by cancer patients and blood donors using the leukocyte apoptosis assay.

METHODS AND MATERIALS

Probes were obtained from 105 healthy blood donors, 48 cancer patients displaying normal sensitivity to radiotherapy, 12 cancer patients displaying hypersensitivity to radiotherapy, 12 Ataxia telangiectasia blood donors, and 4 additional individuals with genetic diseases of potentially modified radiosensitivity; 2 neurofibromatosis (NF) donors, a Nijmegen breakage syndrome (NBS) donor, and an Immunodeficiency, Chromosome fragility, Facial anomaly syndrome (ICF) donor. Heparinized blood was diluted in medium, irradiated, and left to incubate for 48 h. CD4 and CD8 T-lymphocyte DNA was stained with propidium iodide and the cells were analyzed by flow cytometry.

RESULTS

Radiation-induced apoptosis depended on age and cell type. Cohorts of hypersensitive cancer patients, NBS and AT donors displayed compromised apoptotic response. An asymmetric apoptotic response of T-lymphocytes was observed in an ICF donor and a cryptic hypersensitivity donor. Two NF donors displayed no abnormal sensitivity to radiotherapy but compromised apoptotic T-cell response to X-rays.

CONCLUSION

Our studies reveal 4 physiologic sources of variation in radiation response-2 are genetic: cryptic hypersensitivity and hereditary disease, and 2 are epigenetic: cell type and donor age. They emphasize the important role of proteins involved in the recognition and repair of DNA double-strand breaks in determining the response of individuals to radiotherapy.

Breast conservation in the 21st century

As breast cancers are diagnosed at increasingly early stages, and there is little biological rationale for mastectomy in most patients, breast conservation is likely to be practised with increased frequency in the future. Newer breast imaging techniques, particularly magnetic resonance imaging (MRI), should contribute to improved pretherapy planning, both aiding in the selection of patients for conservation approaches, and estimating the residual tumour burden following minimally invasive surgical interventions. Image-guided tumour mapping may permit local treatment to be individualised, most importantly allowing definition of subgroups not requiring treatment directed at the whole breast. Moreover, interventional radiology opens new possibilities for focalised treatments, which may come to be employed in the management of small lesions. The increasing use of primary chemo- or chemoendocrine therapy will also tend to favour the option of breast conservation. Functional imaging techniques, including MRI, may prove valuable in the assessment of response to medical therapy, allowing more individualised use of radiotherapy and surgery. Technical progress and the development of biological response modifiers may further improve the therapeutic ratio associated with radiation treatment.

Repeatability and prognostic impact of the pretreatment pO(2) histography in patients with advanced head and neck cancer

PURPOSE

The purpose of this study was to evaluate the repeatability and the predictive relevance of the pretreatment pO(2) histography on the survival of patients with advanced head and neck cancer.

PATIENTS AND METHODS

From July 1995 to August 1998, polarographic pO(2) measurements of lymph node metastases before therapy were performed in altogether 60 patients with histologically proven squamous cell carcinoma of the head and neck using the Eppendorf histograph. Forty-one of 60 patients were treated with an accelerated-hyperfractionated radiotherapy regimen with or without simultaneous chemotherapy as part of a multicenter phase III study. In 23 of 60 patients, two repeated independent measurements of the same tumor were performed with a time interval of approximately 24 h between the two measurements.

RESULTS

The multivariate analysis revealed the fraction of pO(2) values </=2.5 mmHg as the only significant prognostic factor for the survival (P=0.05) in the 41 study patients. No correlation was found between tumour oxygenation and the volume of the measured lymph node metastases or the haemoglobin concentration. The coefficient of variation of the repeated measurements representing the assay variability was 57-68% of the total variation.

CONCLUSION

Our data support the concept of the relevance of the pretreatment tumour hypoxia for the prognosis of patients with head and neck cancer after fractionated radiotherapy. Because of the relative poor repeatability of the pO(2) histography and the small patient number, further studies are required to confirm this finding and to evaluate the most relevant oxygenation parameter for clinical endpoints.

Altered apoptotic profiles in irradiated patients with increased toxicity

PURPOSE

A retrospective study of radiation-induced apoptosis in CD4 and CD8 T-lymphocytes, from 12 cancer patients who displayed enhanced toxicity to radiation therapy and 9 ataxia telangiectasia patients, was performed to test for altered response compared to healthy blood-donors and normal cancer patients.

METHODS AND MATERIALS

Three milliliters of heparinized blood from each donor was sent via express post to the Paul Scherrer Institute (PSI) for subsequent examination. The blood was diluted 1:10 in RPMI medium, irradiated with 0-, 2-, or 9-Gy X-rays, and incubated for 48 h. CD4 and CD8 T-lymphocytes were then labeled using FITC-conjugated antibodies, erythrocytes were lysed, and the DNA stained with propidium iodide. Subsequently, cells were analyzed using a Becton Dickinson FACScan flow cytometer. Radiation-induced apoptosis was recognized in leukocytes as reduced DNA content attributed to apoptosis-associated changes in chromatin structure. Apoptosis was confirmed by light microscopy, electron microscopy, and by the use of commercially available apoptosis detection kits (in situ nick translation and Annexin V). Data from hypersensitive individuals were compared to a standard database of 105 healthy blood-donors, and a database of 48 cancer patient blood donors who displayed normal toxicity to radiation therapy. To integrate radiosensitivity results from CD4 and CD8 T-lymphocytes after 2 and 9 Gy, z-score analyses were performed.

RESULTS

A cohort of 12 hypersensitive patients was evaluated; 8 showed enhanced early toxicity, 3 showed enhanced late toxicity, and 1 showed both. The cohort displayed less radiation-induced apoptosis (-1.8 sigma) than average age-matched donors. A cohort of 9 ataxia telangiectasia homozygotes displayed even less apoptosis (-3.6 sigma).

CONCLUSION

The leukocyte apoptosis assay appears to be a useful predictor of individuals likely to display increased toxicity to radiation therapy; however, validation of this requires a prospective study.

Programmed cellular response to ionizing radiation damage

Three forms of radiation response were investigated to evaluate the hypothesis that cellular radiation response is the result of active molecular signaling and not simply a passive physicochemical process. The decision whether or not a cell should respond to radiation-induced damage either by induction of rescue systems, e.g. mobilization of repair proteins, or induction of suicide mechanisms, e.g. programmed cell death, appears to be the expression of intricate cellular biochemistry. A cell must recognize damage in its genetic material and then activate the appropriate responses. Cell type is important; the response of a fibroblast to radiation damage is both quantitatively and qualitatively different from that of a lymphocyte. The programmed component of radiation response is significant in radiation oncology and predicted to create unique opportunities for enhanced treatment success.

Acute radiation reaction and local control in breast cancer patients treated with postmastectomy radiotherapy

BACKGROUND

We have investigated the variation of acute radiation reactions in medium-risk patients with postmastectomy radiotherapy with regard to a possible correlation between radiation reaction of normal tissues and local tumor control.

MATERIAL AND METHODS

From 1985 through 1991, a total number of 194 patients received postmastectomy radiotherapy for breast cancer pT1-2pN0-2M0 at the University of Halle-Wittenberg. The lymphatics were irradiated by an anterior 9-MV photon field and the chest wall by an individually shaped anterior field with 9-MV electrons. Both fields received single doses of 2 Gy 5 times weekly up to a total dose of 44 Gy to the chest wall and 50 Gy to the lymphatics. All patients were routinely evaluated once weekly during radiotherapy for acute side effects by one examiner. Skin erythema was classified as mild, moderate or severe, esophagitis as being present in form of dysphagia or not and pneumonitis, if present, as asymptomatic (visible only on repeated chest X-rays) or clinically symptomatic. A differential blood count was also carried out once weekly. For this analysis, the records of all patients were retrospectively reviewed. The median follow-up at the time of analysis was 4.2 years.

RESULTS

Of the patients, 98 (51%) had a mild, 53 (27%) moderate and 43 (22%) a severe erythema. Furthermore, 38 patients (20%) had signs of esophagitis, 13 (7%) had asymptomatic and 26 (13%) symptomatic pneumonitis. Patients with severe erythema or erythema plus esophagitis and pneumonitis had a more pronounced decrease in lymphocyte count during treatment than patients with mild erythema: the lymphocyte nadir was 0.14 vs 0.73 Gpt/l in patients with severe vs mild erythema, and 0.36 vs 0.69 Gpt/l in patients with erythema plus esophagitis plus pneumonitis vs patients with erythema only, p < 0.05. Of the patients, 44 (22%) developed chronic side effects, mostly arm edema. There was no correlation between acute and late effects. An overall number of seven local recurrences (3.6%) occurred. The risk of developing a local recurrence within 5 years after treatment was 0% in patients with severe erythema or erythema plus esophagitis/pneumonitis vs 7% in patients with mild erythema only; this difference was marginally significant, p = 0.055.

CONCLUSIONS

This analysis showed a trend towards better local control in patients with severe acute radiation reaction of normal tissue. The data support a recent publication by Dahl and coworkers showing a linkage between acute radiation reaction of normal tissue and tumor response in patients with preoperative radiotherapy for rectal cancer. The correlation between acute normal tissue reaction and local control might be explained by interindividual variations in the intrinsic, genetically determined radiosensitivity. However, local factors might also be involved, e.g. induction of a cytokin cascade in cases of acute reactions in normal tissues.