The following electronic databases were searched through February 2006: MEDLINE, PreMEDLINE, and the Cochrane Collaboration Library. Results were supplemented with hand searching of selected reviews and personal files. The following MeSH terms and text words were used in a core search: “neoplasms,” “survivors,” “bone marrow diseases,” “antineoplastic agents,” “Herceptin,” “trastuzumab,” “radiotherapy,” “heart diseases,” and “respiratory tract diseases.” “Heart,” “lung,” “mediastinum,” and “cardiovascular system” MeSH terms were paired with the qualifiers “radiation effects” and “drug effects.” Focused searches were done using the term screening and the MeSH subheading “prevention and control.” Due to the expected limited number of randomized controlled trials for screening studies in the target population, study design was not limited to randomized controlled trials, but included any study type, including cohort designs, case series, evaluation studies, and prospective studies. Case reports and non-English language articles were excluded. Letters, commentaries, abstracts not yet published in manuscript form, and editorials were reviewed for any new information. Existing practice guidelines from other organizations were also reviewed.

Articles were originally selected for inclusion in the systematic review of the evidence if they met the following criteria: the study examined a screening intervention for cardiovascular or pulmonary disease; the study population consisted of cancer survivors who had previously received chemotherapy and/or radiotherapy and/or trastuzumab; primary outcomes of interest included screen-detected cardiovascular or pulmonary disorders and the effect of screening on cardiovascular or pulmonary morbidity, quality of life, and overall survival. Primary consideration was given to studies with 50 or more patients and to studies published after 1994; however, studies not meeting these criteria were considered when they provided important additional data or data not available elsewhere. Additional review articles were obtained for reference.

An initial abstract screen was performed by ASCO staff. The ASCO panel reviewed all remaining potentially relevant abstracts identified in the original literature searches to select studies pertinent to its deliberations. Two panel members independently reviewed each abstract for its relevance to the clinical questions, and disagreements were resolved by third-party review. Full-text articles were then reviewed for all selected abstracts. Evidence tables were developed based on selected studies that met the criteria for inclusion.

Originally, the primary outcomes of interest included: the incidence of screen-detected cardiovascular or pulmonary disease; the effect of screening on cardiovascular or pulmonary morbidity; quality of life; and overall survival. Also considered as secondary outcomes were disease-free survival, treatment risks, education or increased awareness, and cost effectiveness. Subsequent to the finding that there were insufficient data in these primary outcome areas, attention was focused on the incidence or prevalence of cardiovascular or pulmonary disease in long-term cancer survivors.

Anthracycline cardiomyopathy is characterized by a dose-dependent progressive decrease in systolic left ventricular function often resulting in CHF. In the adult survivor, it is clinically indistinguishable from CHF due to other causes. Several factors increase the risk and are listed in Table 1. The development of cardiac toxicity is independent, for the most part, of the type of the underlying cancer treated⁵ and strictly related to treatment protocol. There is also evidence that the risk can be reduced but not eliminated with cardioprotective drug use (dexrazoxane)⁶ and with the use of epirubicin and pegylated preparations. It can also be reduced by alterations in administration schedule (eg, once per week v once every 3 weeks) or continuous infusion schedules.

The cardiomyopathy may also present as diastolic dysfunction that is either symptomatic or asymptomatic.⁷ In pediatric patients who received anthracyclines, both systolic and diastolic abnormalities may present with a latency period up to at least 25 years after the completion of therapy and the risk of cardiac failure clearly increases over time.⁵ It has become increasingly recognized that symptomatic CHF may be preceded by asymptomatic or subclinical cardiac dysfunction undetectable at bedside evaluation and/or defined by abnormalities measured by noninvasive imaging techniques (echocardiogram, multigated [MUGA], or radionucleotide angiography) in presumed healthy survivors of cancer.⁸

The risk of developing cardiotoxicity is mainly related to the total cumulative dose of doxorubicin (1% to 5% up to 550 mg/m², 30% at 600 mg/m², and 50% at 1g/m² or higher) with individual variation. The risk increases proportionally to the total accumulated dose in a nonlinear fashion,⁹ so that there probably is no safe dose of doxorubicin.⁵ It is increasingly recognized that abnormalities in noninvasive studies can be found in greater frequency and at a lower cumulative dose than previously reported.^10,11

The prevalence of subclinical cardiotoxicity varies widely between studies with the majority of evidence coming from two separate sources: survivors of pediatric cancer, dominated by long-term follow-up studies of children treated for Hodgkin's disease (HD) and adult studies mainly focused on female survivors of breast cancer. Because the mechanism of disease and the approach to study have been different between the pediatric and adult populations, they will be discussed separately.

Late cardiotoxicity in adults has been studied in breast cancer patients treated with anthracyclines (mainly doxorubicin and epirubicin), acute myeloid leukemia patients treated with idarubicin, or patients treated with doxorubicin for various hematologic malignancies. A total doxorubicin dose of 550 mg/m² has been the traditional cutoff and most studies have looked at populations receiving doses between 450 and 550 mg/m². The incidence of subclinical cardiomyopathy in patients receiving lower doses of doxorubicin has been studied by Hequet et al,¹² and Table 2 ^12-15 summarizes studies (with a variety of study designs) reporting on cardiotoxicity after treatment with anthracyclines.

Beginning at the 5-year mark after treatment completion, the relative risk of cardiotoxicity among patients who received doxorubicin increased compared with untreated patients.¹³ These results emphasize the importance of recognizing the late cardiac effects of treatment especially as the number of long-term survivors treated with anthracyclines continues to increase.

There are very few studies that have examined the risks of anthracyclines and radiation therapy. In one study, breast cancer patients randomly assigned to receive a total cumulative dose of 225 mg/m² with or without radiation had no increase in clinically significant cardiac events relative to the expected rates based on the Framingham heart study, whereas patients who received 450 mg/m² with radiation had significantly higher rates.¹⁶ The concurrent use of doxorubicin and radiation led to a significantly higher rate of CHF in left-sided breast cancers than those with right-sided breast cancers and concurrent use of these two treatments is no longer used in the clinic.¹⁷

After cisplatin-based chemotherapy, survivors of testicular cancer have an increased risk of cardiovascular risk factors and an increased risk of cardiovascular disease compared with an untreated matched population (Table 3). ^18-20 Survivors treated with cisplatin-based chemotherapy have an excess of cardiac risk factors and an increased risk of the development of premature atherosclerosis. No study has systematically evaluated the role of routine noninvasive testing or the role of specific prophylactic treatment in this population.

Abnormal cardiac function occurs with long-term follow-up of patients of all ages who received anthracycline-based chemotherapy for any cancer diagnosis. For purposes of future study, a higher risk population may be defined at the extremes of age, higher cumulative dose, association of mediastinal radiation, and female sex. In contrast with adults, comprehensive serial cardiac monitoring has been performed in pediatric cancer survivors who were treated with anthracyclines. The results show that late symptomatic decompensation is usually preceded by a latent period, often marked by asymptomatic or subclinical LVD. The incidence of abnormal cardiac function appears to increase with the length of follow-up after treatment, and this is particularly true in patients treated with higher total cumulative doses of anthracyclines.

There is a paucity of observational studies that address serial changes in EF in asymptomatic adult survivors, and no studies address cardiac treatment of long-term survivors with asymptomatic LVD. Of the noninvasive studies, echocardiogram may provide the most information in assessing overall cardiac function in this population.

Mediastinal radiation causes inflammation and progressive fibrosis of all of the structures of the heart. Symptomatic heart disease has been well defined in survivors of mediastinal radiation for HD. In a landmark study published in 2003, Heidenreich and associates³⁸ defined the status of asymptomatic valvular and myocardial disease in long-term HD survivors. They prospectively performed echocardiogram screening in 294 asymptomatic patients (mean age, 42 ± 9 years) treated with mediastinal radiation (mean dose, 43 ± 3 Gy) who ranged from 2 to 33 years (mean, 15 years) from treatment completion and separated these patients into those surviving 2 to 10 years, 11 to 20 years, and longer than 20 years from treatment. They found frequent abnormalities in valvular function, left ventricular mass, and systolic function in this population. The prevalence of these abnormalities was compared to the Framingham population and was several-fold higher. All abnormalities increased over time elapsed from treatment completion. The most common valvular abnormality was aortic regurgitation, with less frequent involvement of the mitral and tricuspid valves. Sixty percent of patients that were followed in excess of 20 years had at least mild aortic insufficiency that would meet the criteria for endocarditis prophylaxis. Of interest, only 5% of these patients were recognized to have aortic insufficiency by physical examination performed by an experienced cardiologist. In addition to valvular disease, 36% of these patients had abnormal systolic function marked by a decreased fractional shortening. The development of valve disease is similar to the general population regarding hemodynamics, natural history, and progression.

The same group subsequently reported on diastolic function of the 294 prospectively screened asymptomatic patients.³⁹ They found a high prevalence of diastolic dysfunction in this asymptomatic group, with increased incidence related to older age, the presence of hypertension, diabetes, wall motion abnormalities, and those with a longer latency period from radiation treatment to screening. For patients who had completed radiation therapy for 11 to 20 years and longer than 20 years, 15% and 23%, respectively, had mild to moderate diastolic dysfunction. Diastolic dysfunction was seven-fold more common in this population than in community data from Rochester, MN. Moreover, on stress echocardiography, CAD was found to be significantly more common in patients with diastolic dysfunction than in patients with normal function (10% v 2%; P = .005). Deaths or events due to CAD were also significantly more common in patients with diastolic dysfunction compared with patients with normal function (25% v 8%; P = .002).

Multiple studies have shown that HD survivors treated with mediastinal radiation are at increased risk for fatal cardiovascular disease, with relative risk ranging from 2.2 to 7.2 compared with age- and sex-matched general population controls.^41-43 This risk becomes manifest 5 to 10 years after therapy completion^44,45 and cardiovascular disease remains a major cause of morbidity and mortality after treatment for HD. The clinical presentation of radiation-induced CAD is similar to CAD in the general population. It may be silent, present with angina, or cause sudden death. Silent myocardial infarction may be more common due to the possible damage to nerve endings in and around the heart secondary to radiation. When CAD is due to radiation therapy, it more commonly affects the right coronary artery, the left anterior descending coronary artery, and the left main coronary artery, with relative sparing of the left circumflex system. A characteristic of radiation-induced disease is the proximal and often ostial distribution of disease.³⁷

Radiation-induced cardiovascular disease is responsible for one fourth of the deaths not directly attributable to HD itself with an increase in relative risk of 2.2 to 3.1. The risk is associated with treatment total more than 40 Gy. The relative risk for myocardial infarction is 4.2 and 6.7 for myocardial infarction or sudden death.³⁵ Hull et al³⁷ found a 10.4% incidence of CAD at a median follow-up of 11.2 years post-treatment in a survey of 415 patients. The results of at least one study³⁷ suggest that the major modifier in the development of ischemic heart disease is the presence of traditional cardiovascular risk factors³⁷ so that, at a minimum, a history of mediastinal radiation therapy could logically be added to that list.

The American College of Radiology⁴⁶ has developed appropriateness criteria for routine follow-up evaluation of HD survivors who are asymptomatic. Their expert panel recommended routine exercise tolerance testing and echocardiography in symptomatic patients and periodic screening in patients depending on the mediastinal radiation dose, cumulative anthracycline dose, and presence of other cardiac risk factors.

The data are similar for pediatric survivors of HD treated with radiation therapy alone or combined modalities. These are important because of the large potential number of adolescents/young adults who will transition to adult oncologists for long-term follow-up. A published review of 14 articles found that the cumulative incidence of clinical CVD assessed in nine studies ranged from 0.3% to 22.8%.⁴⁷ In a prospective cardiac screening study, Adams et al⁴⁰ reported on the incidence of asymptomatic cardiac disease in survivors of childhood HD who were 5.9 to 27.5 years post-treatment (median, 14.3 years) that consisted of a median of 40 Gy (range, 27 to 51.7 Gy) at a median age of 16.5 years (range, 6.4 to 25 years). They found that 42% had significant valve defects, 75% had conduction defects, and 22% had echocardiographic changes suggestive of restrictive cardiomyopathy.

Radiation therapy is an integral component of the locoregional management of breast cancer and has efficacy in reducing local recurrence. Breast and chest wall radiation treatment have been associated with an increase in cardiac toxicity, mainly manifested by ischemic heart disease with angina, myocardial infarction, or sudden death.

The early literature focused on women who were treated after mastectomy with adjuvant radiation therapy. Those with left-sided breast cancer were shown to have an increased incidence of fatal CVD that manifested with prolonged follow-up.^48-51 In studies published before 1990, reflecting premodern radiation oncology, the risk of radiation-induced CAD was thought to be increased in patients with left-sided breast cancer compared with right-sided, with the excess mortality balancing or exceeding the risk reduction gained by adjuvant radiation therapy. More recent meta-analyses have shown a survival benefit for radiation treatment after surgery (mainly breast conserving), and with modern techniques the incidence of cardiac disease for left and right breast radiation therapy is similar with an overall risk reduction compared with no radiation treatment.^52-56 Whether or not there is a significant laterality effect, there is no question that coupled with risk factors for cardiovascular disease, radiation therapy of either side increases the risk of future events.

Data from the SEER database⁵⁴ on 8,363 patients with left breast cancer and 7,907 patients with right breast cancer treated from 1986 to 1993 with adjuvant radiation therapy at a mean follow-up time of 9.5 years (0 to 15 years) showed no laterality differences for the development of ischemic heart disease (9.8%), valvular heart disease (2.9%), and conduction disease (9.7%). Another study using SEER data⁵⁵ examined the risk of cardiac death in 27,283 women in three periods of radiation therapy that reflect the transition to modern radiotherapy: 1973 to 1979, 1980 to 1984, and 1985 to 1989. This study confirmed the lack of laterality in the incidence of ischemic heart disease that grew from changing radiation treatment delivery technique and, of equal importance, showed that the risk of death substantially decreased over time from 1973 to 1979 with the 15-year post-treatment risk of cardiovascular death decreased from approximately 13% in 1973 to 1979 to 5.5% in 1985 to 1989.

Radiation-induced late pericardial disease (months to years after treatment) may be silent, with the incidental discovery of asymptomatic pericardial effusions, or may present with hemodynamic compromise secondary to a reduction in ventricular filling and cardiac output. The latter can be due just to excess pericardial fluid with tamponade, purely constrictive without pericardial fluid, or a combination of both: effusive constrictive pericarditis. All of the latter are symptomatic, not silent, and present with typical signs and symptoms. There is no evidence that interventions can alter the course of hemodynamically inconsequential and clinically silent effusions.

Approximately 20% to 25% of those with late pericarditis progress to develop chronic constrictive disease or acute tamponade 5 to 10 years post-therapy. The incidence of pericardial disease decreased from 20% to 2.5% at one center with the use of modern techniques.⁵⁷

Radiation-induced myocardial disease differs from the predominantly systolic dysfunction associated with anthracyclines. Instead, it presents with diastolic disease and restrictive hemodynamics (in the absence of combined treatment with an anthracycline). Modern techniques have reduced the risk of systolic dysfunction but have not changed the course of restrictive disease.⁵⁸ In a series of 21 asymptomatic survivors⁵⁹ treated with 20 to 76 Gy (mean, 35.9 Gy) for HD before 1983, 57% had an abnormal EF by radionucleotide angiography 7 to 20 years after treatment (mean, 14.1 years) compared with a modern technique series⁶⁰ evaluating 50 HD survivors age 18.5 to 47.5 years (mean, 35.1 years), 1 to 30 years after treatment (mean, 14.1 years) in which 4% had an abnormal EF by radionucleotide angiography and 16% had an abnormal filling rate consistent with diastolic dysfunction.

Life-threatening arrhythmias and conduction disease occur years after treatment. There is a wide spectrum of abnormalities that include the sick sinus syndrome, all forms of atrio-ventricular block, and bundle branch block. All are easily identified by a routine ECG.

The frequency of serious conduction abnormalities in long-term asymptomatic cancer survivors ascribed solely to mediastinal radiation therapy is not known and is probably overstated in the literature since disturbances may not occur for years after treatment cessation, making causality a difficult assumption. There are few prospective studies reporting incidence.

Carotid artery stenosis is a recognized late treatment effect of radiation therapy for head and neck tumors.^61-63 Hull³⁷ reported an incidence of carotid and/or subclavian artery disease of 7.4% at a mean of 17 years post-treatment cessation. This complication impacts the risk of stroke with reported actuarial stroke rate of 12% at 5 years⁶⁴ and a 10- and 15-year relative risk of 10.1% (95% CI, 4.4% to 20%) and 12.0% (95% CI, 6.5% to 21.4%),⁵⁴ respectively. Wai-man Lam⁶⁵ found greater than 50% stenosis in the extracranial carotid artery in 24 of 80 patients. Nine patients had historical evidence of a cerebrovascular accident or transient ischemic attack and 15 patients were asymptomatic. Only seven patients had a carotid bruit recognized by auscultation and no control patient had any stenosis more than 50%. Cheng et al⁶⁶ performed serial carotid ultrasound studies on 95 patients postexternal neck radiation therapy and found an accelerated progression rate from less than 50% at baseline over a 3-year period of 15.4% compared with a control group (nonirradiated) rate of increase of 4.8%.

Knowledge about the increased risk of carotid disease and its progression is important since there is definitive surgery that has been as successful in radiation-induced disease as in atherosclerotic disease^67-71 and has been shown to reduce the incidence of transient ischemic attack and stroke.

Trastuzumab improves the outcome of women with HER-2/neu amplified and/or overexpressing breast cancers in the metastatic and adjuvant treatment settings (Table 7). ^68-71,79 The significant improvements in disease-free, progression-free, and overall survival are accompanied by a small increased risk of symptomatic CHF and a higher risk of asymptomatic declines in LEVF.⁸⁰ Although there are numerous phase II trials of trastuzumab as single agent or in combination with chemotherapy, the best sources for trastuzumab-related cardiotoxicity are the four prospective randomized controlled trials in the adjuvant setting (Table 7). In each of these trials, the end points reported were cardiac deaths; symptomatic CHF primary defined as New York Heart Association (NYHA) class III (marked limitation of physical activity comfortable at rest, but less than ordinary activity causes fatigue, palpitation, or dyspnea) and class IV (unable to carry out any physical activity without discomfort, symptoms of cardiac insufficiency at rest; if any physical activity is undertaken, discomfort is increased); or asymptomatic declines in LVEF. In the pivotal trial of chemotherapy with or without trastuzumab in metastatic disease,⁷⁴ prospective measurements of cardiac function were not performed, but an independent cardiac evaluation committee retrospectively adjudicated the cases of cardiac dysfunction. Concurrent administration of trastuzumab with doxorubicin and cyclophosphamide (AC) resulted in a 16% incidence of NYHA classes III and IV relative to 3% with AC alone, and this combination is no longer used in the clinic. Preliminary experience with other anthracyclines, such as epirubicin or liposomal encapsulated doxorubicin and trastuzumab, is limited to small phase I and II trials^81,82 in which the observed incidence of NYHA classes III and IV is lower. In contrast, paclitaxel or docetaxel and trastuzumab combinations⁸³ result in 1% to 2% incidence of NYHA class III or IV relative to ≤ 1% with taxane alone, and taxane/trastuzumab alone or after AC has rapidly become the new standard for metastatic and adjuvant treatment, respectively.

Based on the incidence of CHF observed in the pivotal trial, the adjuvant trials included serial prospective measurements of LVEF by MUGA and/or echocardiogram performed pretreatment, after anthracycline-based chemotherapy, and typically every 6 to 12 months after randomization.^75-77,79 In addition, patients with pre-existing heart diseases such as CHF, myocardial infarction, angina requiring medications, or arrhythmia requiring medications were excluded from the trials. The criterion for permanently stopping trastuzumab was the development of symptomatic cardiac dysfunction. The criteria for holding trastuzumab were based on LVEF decrements in asymptomatic patients with a repeat measurement of LVEF after 4 weeks. If there was no improvement in LVEF, trastuzumab was permanently discontinued.

In the combined analysis of the Intergroup and National Surgical Breast and Bowel Project (NSABP) trial B-31, LVEF was evaluated before and after AC, before receiving paclitaxel/trastuzumab.^75,78,84 Only those patients with LVEF above the lower limit of the normal range or in whom it had decreased less than 16% from the baseline value were eligible to receive trastuzumab. Ninety-three percent of patients received trastuzumab after AC. Overall, the incidence of NYHA classes III and IV for the trastuzumab-containing treatment group was 4.1% (NSABP) and 2.9% (Intergroup) versus 0.8% and 0% in the respective control groups.

A more detailed analysis of adverse cardiac events (symptomatic and asymptomatic) in NSABP B-31 was recently published.⁷⁸ A cardiac event was defined as NYHA class III or IV CHF and cardiac death. The cumulative incidence of cardiac events at 3 years was 4.1% versus 0.8% (hazard ratio [HR], 5.9; 95% CI, 2.3 to 15.3; P = .0001) in the trastuzumab and control groups, respectively. There was one cardiac death in the entire cohort and that patient was in the control group. In a multivariate analysis, increasing age and decreasing LVEF after AC were independent risk factors for CHF.

In the Breast Cancer International Research Group (BCIRG) 006 trial, patients were randomly assigned to a nonanthracycline trastuzumab treatment (docetaxel, carboplatin, and trastuzumab) versus AC followed by docetaxel with or without trastuzumab.⁷⁷ Symptomatic cardiac events were defined as CHF; grades 3 or 4 ischemia/infarction and arrhythmia; and asymptomatic declines in LVEF of more than 15% or below the lower limit of normal. At a median follow-up of 23 months, the incidence of symptomatic cardiac events was 2.3% versus 1.2% (P = .046) for AC followed by docetaxel with or without trastuzumab, respectively; in the nonanthracycline-containing trastuzumab arm, it was 1.2% (P = .99 relative to the control arm). Likewise, asymptomatic declines in LVEF were 2.4% versus 0.6% (AC followed by docetaxel with or without trastuzumab, respectively P = .001) and 0.6% v 0.4% (relative to nonanthracycline trastuzumab arm, P = .54).

Unlike the Intergroup, NSABP, and BCIRG trials, patients in the Herceptin Adjuvant Trial (HERA) trial received trastuzumab after completion of adjuvant chemotherapy.⁷⁶ Ninety-four percent of patients received prior anthracyclines with or without taxanes, and approximately 50% of these patients received epirubicin. Using a more restricted definition of cardiac events, the overall incidence of symptomatic CHF (including NYHA class III or IV) or decrease from baseline of more than 10% below 50% was 1.73% v 0.06% (P ≤ .001) for trastuzumab and control, respectively. Asymptomatic declines in LVEF were observed in 7.1% v 2.2% (P < .001). The HERA trial included a second random assignment to 2 years versus 1 year of trastuzumab; however, no cardiac safety or efficacy results are available.

A recent subset analysis of a small number of HER-2–overexpressing breast cancers showed that a 9-week duration of trastuzumab with either adjuvant docetaxel or vinorelbine followed by epirubicin-containing chemotherapy reduced the HR for disease-free survival and was not associated with any cardiac dysfunction.⁷⁹ To our knowledge, this is the first such trial to suggest that a short duration of trastuzumab may be effective. However, larger trials are needed to confirm this result.

The clinical course of trastuzumab-related CHF differs from anthracycline-induced CHF in several important ways: it very rarely causes death; does not cause myocyte drop-out or other changes observed in electron microscopy of endomyocardial biopsies typical of anthracycline-induced CHF; and is reversible with/or without medical treatment.⁷³ In addition, there is evidence that cumulative dose and duration of trastuzumab do not contribute to increasing the risk of CHF.⁸⁵ In fact, Ewer et al⁷² have proposed a new category of “Type II myocardial dysfunction” to describe trastuzumab-related CHF as opposed to “Type I myocardial damage” that results from anthracyclines. Based on the available experience, the majority of patients who developed trastuzumab-related CHF and discontinued trastuzumab had recovery of LVEF with subsequent follow-up.

The optimal duration, frequency, and method of cardiac monitoring during trastuzumab treatment remain unknown. Criteria for discontinuing anthracyclines on the basis of LVEF decrements evaluated by resting MUGA scans or echocardiograms were developed from retrospective studies. One small prospective blinded study of serial LVEF for cardiac monitoring in patients treated with epirubicin concluded that MUGA scans were too insensitive to predict CHF.⁸⁶ The same approach used for anthracycline cardiac monitoring with serial assessments has been applied in the adjuvant trastuzumab trials. Using cut-offs for transiently stopping or permanently discontinuing trastuzumab based on LVEF decrements from baseline, the incidence of CHF NYHA classes III and IV ranges from 3% to 4%. Plasma markers of cardiac dysfunction, such as cardiac troponin or BNP, have been collected in some adjuvant trastuzumab trials, but no results are available.

Although many chemotherapeutic agents may cause pulmonary toxicity (Table 8), bleomycin is the most studied and is the focus of this section of the review. Bleomycin is commonly used in the treatment of GCTs and malignant lymphoma, especially HD. Bleomycin-induced pneumonitis (BIP) is the most common form of bleomycin-related pulmonary toxicity. Risk factors that predispose to BIP include the cumulative dose of bleomycin, the patient's age, smoking, renal dysfunction, mediastinal radiation therapy, and administration of oxygen.^87,88 The incidence of BIP ranges from 0% to 46%, depending on the patient population being studied and the criteria used to diagnose this entity.⁸⁷

GCT patients receiving bleomycin-containing chemotherapy are an ideal population for examining the incidence of pulmonary toxicity. These patients are generally younger and do not receive mediastinal radiation therapy which is also associated with pulmonary toxicity (Table 9). ^89-93 As noted, in most GCT studies the incidence of clinically-significant BIP is quite low. In a large trial comparing mechlorethamine, vincristine, procarbazine, prednisone, doxorubicin, bleomycin, and vinblastine (MOPP/ABV) with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) in HD patients, acute pulmonary toxicity occurred in 25% to 31% of patients. However, unlike the GCT population, approximately 10% of these patients in this study also received radiation therapy.

Radiation pneumonitis has been reported in patients who have undergone mediastinal radiation therapy for lung cancer, HD, breast cancer, and other cancers that require radiation therapy to the thorax. Radiation pneumonitis is reported in 5% to 15% of patients receiving definitive external-beam radiation therapy for lung cancer.⁹⁴ Clinical factors that may increase the risk of radiation pneumonitis include concomitant chemotherapy, previous irradiation, and recent withdrawal of steroids.⁹⁴ A retrospective analysis of 1,911 patients who underwent combined-modality therapy for lung cancer demonstrated that the overall risk of radiation pneumonitis was 7.8%.⁹⁵ In a multivariate analysis, the daily fraction dose, number of daily fractions, and total dose were associated with radiation pneumonitis (P < .001; P < .018; and P < .003, respectively).

The incidence of radiation pneumonitis is lower in HD and breast cancer patients who received mediastinal radiation therapy compared with lung cancer patients. A series of 590 patients with stage IA-IIIB HD who received mantle radiation therapy alone demonstrated a 3% risk of radiation pneumonitis with radiation therapy alone. The risk of radiation pneumonitis was significantly increased to 11% when radiation therapy was combined with chemotherapy (P = .0001).⁹⁶ Patients with breast cancer who undergo breast radiation therapy as part of a breast conserving approach have less than 1% risk of radiation pneumonitis.⁹⁷

Pulmonary complications occur frequently after BMT. In one study of 311 patients with hematologic malignancies who underwent autologous, allogeneic, or syngeneic BMT, the 2-year risk of interstitial pneumonitis was 26.8%.⁹⁸ In 90% of these cases, cytomegalovirus was detected or the cases were considered idiopathic. Patients who underwent allogeneic BMT had the highest incidence of interstitial pneumonitis (34.1%). In patients who underwent allogeneic stem-cell transplantation, age older than 20 years, a history of chronic myelogenous leukemia, alloimmunized donor, prior splenectomy, and graft-versus-host disease were risk factors for the development of interstitial pneumonitis. Interstitial pneumonitis occurs more frequently after busulfan-conditioning regimens as compared with total-body irradiation. The use of lung compensators during total-body irradiation decreases the risk of interstitial pneumonitis.⁹⁹ Patients who receive BCNU as a part of augmented preparative regimens for high-dose chemotherapy and autologous peripheral blood progenitor cell transplantation are at risk for pulmonary toxicity. Up to 30% of these patients develop noninfectious pulmonary complications. BCNU pneumonitis may result in long-term pulmonary function testing abnormalities. In addition, patients may respond to corticosteroids.¹⁰⁰

BIP usually starts insidiously during treatment, but the development of BIP up to 2 years after discontinuation has been reported.¹⁰¹ Most patients who experience BIP will recover with discontinuation of the agent and/or corticosteroid treatment. A minority will progress to pulmonary fibrosis.⁸⁷ The majority of patients who develop BCNU pneumonitis will respond to corticosteroid therapy.¹⁰⁰ Radiation pneumonitis is typically a delayed acute reaction, usually occurring 1 to 3 months after completing mediastinal radiation therapy.⁹⁴ Patients who experience acute radiation pneumonitis will often have a self-limited course, with complete resolution of this process. However, a minority of patients may develop progressive pulmonary fibrosis, usually 6 to 24 months after treatment. Late complications of pulmonary fibrosis include cor pulmonale and respiratory failure. Late pulmonary complications in BMT patients who develop interstitial pneumonitis include the idiopathic pneumonia syndrome and bronchiolitis obliterans.

Pulmonary function testing includes measurement of lung volumes and oxygen diffusion capacity. Pulmonary function test monitoring has been reported in several populations of long-term (> 24 months) cancer survivors treated with chemotherapy and/or radiation therapy (Table 10). ^102-110 Most of the studies have focused on HD survivors. Of interest, in many of these studies, early decline in pulmonary function testing is followed by subsequent improvement over time in most patients.