Unveiling the diagnostic enigma of D‐dimer testing in cancer patients: Current evidence and areas of application

Cancer is a well‐known risk factor for venous thromboembolism (VTE). A combined strategy of D‐dimer testing and clinical pre‐test probability is usually used to exclude VTE. However, its effectiveness is diminished in cancer patients due to reduced specificity, ultimately leading to a decreased clinical utility. This review article seeks to provide a comprehensive summary of how to interpret D‐dimer testing in cancer patients.


| INTRODUCTION
The imbalance of physiological haemostasis can result in disturbances in the complex interplay between platelets, coagulation proteins and endothelial cells. These disturbances can lead to either a prothrombotic state, which can cause thrombus formation at the site of violated vessel integrity or an increased risk of bleeding. 1 After a thrombotic event, the body initiates fibrinolytic processes to maintain a patent vascular system. These processes involve several enzymes that work together to convert insoluble fibrin to soluble fibrinogen degradation products (FDPs). When fibrinogen is converted into fibrin by enzymatic cleavage of the fibrinopeptides A and B, the resulting fibrin monomers can aggregate and align in response to activated factor XIII, calcium and platelets. 2,3 Activated plasmin then cleaves the cross-linked fibrin into soluble FDPs, including the dimeric D-domain (also known as 'D-dimer') as the terminal product 2,3 ( Figure 1).
Representing products of degraded cross-linked fibrin, D-dimers are indicative of in vivo fibrin formation. The evaluation of D-dimers as biomarkers in clinical decisionmaking dates back several decades and is currently regarded as the biochemical gold standard for the laboratory exclusion of clinically suspected VTE. It may also serve as a valuable tool for the monitoring of disseminated intravascular coagulation. 2 An association between cancer and the haemostatic system has been observed in various preclinical and clinical studies. [4][5][6][7] Tumour cells inherit strong procoagulant effects that induce the formation of platelet aggregates, promote metastatic spread and advance tumour progression. 8 Cancer patients face a risk of developing VTE that is 4 to 7 times higher than the general population. 9 The cumulative incidence of VTE, including pulmonary embolism (PE), among cancer patients exhibits significant variation and is influenced by several factors, such as the type of cancer. 10 The extent of coagulation activity, as reflected by D-dimer plasma concentrations, has been linked to an advanced tumour stage and poor outcomes in studies involving patients with different types of cancer. [11][12][13][14] This review article aims to provide a comprehensive summary of the utilization and interpretation of D-dimer testing in patients with diverse types of cancer. It is important to note that, currently, there are no existing guideline recommendations specifically tailored for this particular patient population.

| METHODS
This review article was designed to systematically analyse the diagnostic and prognostic value of D-dimer testing in patients with cancer. The study was conducted based on the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines and conforms to broad EQUATOR guidelines 15,16 (Figure 2). The literature search was performed in the PubMed and Cochrane databases, using search terms such as "D-dimer", "cancer", "thrombosis", "pulmonary embolism", "venous thromboembolism", and "deep vein thrombosis". Additionally, we manually searched for appropriate hits in the reference lists of eligible studies to identify other relevant publications. The inclusion criteria for study selection were articles in English or German language and articles relevant to the current topic of this review.

| Structure of D-dimers
D-dimers are soluble FDPs composed of two crosslinked D fragments of the fibrin protein. This structure allows them to bind to monoclonal antibodies and to be detected in blood as small protein fragments present after the degradation of blood clots during fibrinolysis. 3 Their half-life period is approximately 8 h and they become detectable in the blood approximately 2 h after thrombus formation. 3 Since D-dimers can only arise during the formation and degradation of crosslinked fibrin, they function as a global marker of activation of both the coagulation and fibrinolytic systems. Therefore, they can be used as an indirect marker of thrombotic activity. 2 consider the variations in assay performance and the specific test characteristics of their institution.

Conclusions:
Standardizing D-dimer assays and developing modified pretest probability models specifically for cancer patients, along with adjusted cut-off values for D-dimer testing, could significantly enhance the accuracy and effectiveness of VTE diagnosis in this population.

K E Y W O R D S
cancer, D-dimer, deep vein thrombosis, pulmonary embolism, venous thromboembolism 2.2 | Illustration of different assays for D-dimer measurement D-dimers can be detected and quantified either in whole blood, plasma or serum using monoclonal antibodies that recognize a specific epitope on cross-linked D-dimer molecules, which is otherwise absent on non-cross-linked fibrin and FDPs. 2,17,18 These monoclonal antibodies facilitate the accurate detection and quantification of D-dimers in various laboratory tests.
Over the past decade, several different D-dimer assays have been developed to detect plasmin-mediated FDPs containing cross-linked D fragments. 19 Therefore, physicians should be aware of the diagnostic properties of the specific assay used in their clinic. Generally, there are three types of D-dimer assays: whole-blood agglutination assays, enzyme-linked immunosorbent assays (ELISA) or immunofluorescent assays (ELFA) and latex agglutination assays 17,[19][20][21] (Figure 3). While ELISAs detect D-dimers by capturing and labelling antibodies, whole-blood agglutination assays and latex agglutination assays use a bispecific antibody conjugating with binding sites for both D-dimer and red cell antigens or latex particles. 17,21 Each of these test methods has its specific advantages and limitations. 3,17,20 Table 1 provides a comprehensive summary of the features of the different assay types. Currently, it is not possible to compare the results of D-dimer testing across different assays, even if they were tested with similar formats or techniques. 3,20 Several factors can contribute to the lack of comparability between D-dimer test results, including the use of monoclonal antibodies with varying specificities for Ddimers, differences in assay methodology or instrumentation and variations in the values used to differentiate between positive and negative test results. Moreover, some antibodies may cross-react with non-cross-linked degradation products of fibrin or fibrinogen if the assay antibodies are not specific to the D-dimer fragment. In addition, D-dimer levels may be reported either as Ddimer concentration or as fibrinogen equivalent. To convert fibrinogen equivalent units to D-dimer concentration, the level needs to be halved. 2,3 According to the ESC guidelines, it is recommended to utilize a highly sensitive assay for plasma D-dimer measurement in outpatients and emergency department patients with low or intermediate clinical probability or PE-unlikely. 22 On the contrary, there exist qualitative and quantitative D-dimer tests. Qualitative tests provide the advantage of facilitating immediate diagnostic decisions on the basis of a positive or negative test result without relying on a central laboratory. Research has demonstrated the safety of utilizing qualitative D-dimer tests in primary care settings F I G U R E 1 Composition of the D-dimer molecule. This figure illustrates the process of D-dimer formation. When blood clots, thrombin is produced, which cuts off fibrinopeptide A (FpA) and fibrinopeptide (FpB) from fibrinogen. Fibrinogen is a protein composed of three pairs of polypeptide chains linked together by disulfide bonds to form three globular domains, each consisting of a central E domain and two D-domains on either side. The resulting soluble fibrin monomers polymerize into an insoluble fibrin network, which is further stabilized by activated factor XIII (FXIIIa), creating covalent cross-links between the fibrin molecules. Plasmin, produced during fibrinolysis, degrades fibrinogen and fibrin; however, it is unable to break the covalent bonds between D-domains. FpA, fibrinopeptide A; FpB, fibrinopeptide; FXIIIa, activated factor XIII.
to exclude both deep vein thrombosis (DVT) and PE. 23 However, it is important to note that laboratory-based quantitative assays typically exhibit higher sensitivity when compared to qualitative tests. This higher sensitivity may enhance the reliability of excluding PE, potentially contributing to a safer diagnostic approach. 24 F I G U R E 2 PRISMA flow diagram on the decision-making process for including studies following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Adapted from reference. 15

F I G U R E 3 Diagnostic assays for D-dimer quantification. (A)
Whole-blood agglutination assays use a bispecific antibody conjugate with binding sites for both D-dimer and red cell antigens. (B) Enzyme-linked immunosorbent assays (ELISA) detect D-dimers by capturing and labelling antibodies. (C) The latex agglutination assay uses a bispecific antibody conjugate with binding sites for both D-dimer and latex particles. ELFA, enzyme-linked fluorescence assay; ELISA, enzyme-linked immunosorbent assay.

| Guideline recommendations
Since its introduction in the 1990s, D-dimer testing has become an important diagnostic tool in individuals with suspected VTE. 22,25 Several guidelines recommend Ddimer testing as the first diagnostic step in patients with a low clinical pre-test probability. 22,26 Although being highly sensitive, the low specificity of D-dimer testing partly explains its primary application in outpatients where the pre-test probability is low. 22 While D-dimer values below an assay-specific cut-off point practically rule out the presence of VTE, a positive result cannot reliably differentiate between thrombosis and many other conditions with elevated D-dimers, including cancer, pregnancy, infection, laboratory assessment or a postsurgical state. Conditions and diseases which are associated with elevated D-dimer levels are listed in Table 2. 3 In addition, other confounding factors must be considered when interpreting the test results. Clinical and analytical reasons for false positive results are summarized in Table S1. 27,28 Therefore, D-dimer testing is still discouraged by current guidelines in patients with cancer due to the lack of specificity. [29][30][31] In this context, the number of tests needed to diagnose a case varies, ranging from 2 in patients with low clinical probability to more than 10 in those with the specific situations mentioned above. 22,32 Besides cancer as one of the most important confounding factors, pregnancy represents another highly challenging condition for the correct diagnosis of VTE. 33 Since D-dimer concentrations continuously increase during pregnancy, approximately one-quarter of pregnant women show D-dimer levels above the ruleout cut-off in the third trimester. 22 Nevertheless, Ddimer testing facilitated the exclusion of PE in 12% of women (46 of 395 included patients) based on clinical pre-test probability and laboratory assessment alone. 33 Recommendations for the use of D-dimer testing are summarized in Table 3. It is important to highlight that the current guideline recommendations primarily pertain to patients without cancer. Currently, there are no existing guideline recommendations specifically tailored to address the unique considerations and needs of cancer patients.

| Clinical prediction rules and their application in cancer patients
The combination of a negative D-dimer test result and the Wells score represents an established and validated tool for the exclusion of DVT and PE for most patients. 22,34 Similar to patients without cancer, the combination of a low or unlikely pre-test probability and a negative D-dimer result can effectively rule out the presence of VTE in cancer patients. However, unlike patients without cancer, this combination is relatively infrequent in cancer patients, resulting in the exclusion of only a small number of individuals. Consequently, the clinical usefulness of this approach is limited and not well established. 30,34 The Wells scoring system incorporates malignancy as a single variable but does not take into account specific risk factors that are unique to cancer patients, such as chemotherapy or central T A B L E 1 Illustration of D-dimer assay types and their diagnostic properties.

Latex agglutination assays
Whole-blood agglutination assays venous catheter placement. 35,36 Considering the more frequently observed clinical signs and symptoms that indicate VTE along with the point score given for cancer, current clinical prediction rules tend to categorize cancer patients into higher-risk groups. 30 Therefore, the Wells score has a lower discriminatory power for detecting VTE in cancer patients as opposed to those without cancer. 37 To summarize, the combined approach of using the Wells clinical decision rule and D-dimer testing is less effective for making a diagnosis in cancer patients with suspected VTE compared to those without cancer. 37 Therefore, it is important to develop and prospectively validate clinical prediction rules that are specific to cancer patients to improve discriminative performance. 36

| Proposed clinical decision rule using ROC-optimized D-dimer cut-off values
Numerous combinations of clinical evaluations, plasma D-dimer measurements and imaging tests have been suggested and verified for diagnosing VTE. The standard diagnostic algorithms for suspected VTE are well established. 22 However, they are less effective in cancer patients due to the lack of D-dimer specificity. 29,30 To address this issue, a clinical decision rule with optimized D-dimer cutoff values without haemodynamic instability has been developed and presented in Figure 4. It is important to note that the diagnostic approach may differ depending on the availability and expertise in specific tests across different hospitals and clinical settings.

| Prevalence of VTE in different types of cancer
Armand Trousseau was one of the first to report on the association between cancer and the development of thrombosis in 1865. 11 The prevalence of VTE in cancer patients is estimated to be around 4% to 20%, with the highest rates observed shortly after the initial diagnosis. 38 Overall, cancer patients have an up to sevenfold elevated risk of developing a venous thrombosis compared to the general population, representing one of the most frequent complications and leading causes of death. 39,40 In most cancer entities, the incidence of VTE increases from localized to metastasized cancer. 41 Ten years after diagnosis, the risk of VTE almost equals that of the general population. 11 Different types of cancer have a dissimilar risk of developing a VTE, depending on the tumour biology and the aggressiveness of tumour invasion. 14,42 Several studies have shown that patients with lung or pancreatic cancer are prone to a much higher risk of developing a VTE compared to other entities. 11,14,36 On the contrary, breast and T A B L E 2 Conditions and diseases associated with elevated D-dimer levels. In non-cancer patients • Plasma D-dimer measurement, preferably using a highly sensitive assay, is recommended in outpatients with low or intermediate clinical probability or PE-unlikely to reduce the need for unnecessary imaging and irradiation (LOE IA). • As an alternative to the fixed D-dimer cut-off, a negative D-dimer test using an age-adjusted cut-off (age × 10 mg/L, in patients aged >50 years) should be considered for excluding PE in patients with low or intermediate clinical probability, or those that are PE-unlikely (LOE IIaB). • As an alternative to the fixed or age-adjusted D-dimer cutoff, D-dimer levels adapted to clinical probability should be considered to exclude PE (LOE IIaB). • D-dimer testing is not recommended in patients with high clinical probability, even if using a highly sensitive assay (LOE IIIA). In cancer patients • A negative D-dimer test has the same diagnostic value as in non-cancer patients. • Age-adjusted cut-offs may increase the rule-out of a significant proportion of cancer patients with a low rate of false-negatives. • Edoxaban or rivaroxaban should be considered as an alternative to LMWH, with the exception of patients with gastrointestinal cancer due to the increased bleeding risk with NOACs (LOE IIA).
prostate cancer have a significantly lower chance of developing a VTE. 31,43 D-dimer blood levels in patients with malignant tumours have been shown to be higher compared to patients with benign tumours. [44][45][46][47][48] For example, in patients with ovarian cancer, the peripheral D-dimer blood levels were found to be 17-times higher compared to those with benign ovarian disease, and also in breast cancer patients, the median plasma D-dimer levels were found to be significantly higher in patients with invasive carcinoma compared to patients with either benign breast disease or carcinoma-insitu. 44,45 Regarding pancreatic cancer, a study found more than 300% increased D-dimer values in portal blood compared to patients with chronic pancreatitis. 49 Furthermore, D-dimer testing may also be a useful indicator for distinguishing between lipoma and well-differentiated liposarcoma before surgery. 50 Interestingly, no difference has been reported between healthy patients and patients with benign neoplasia. 44 Herein, D-dimer concentrations of patients with benign ovarian tumours differed not compared with healthy controls. 44 To summarize, D-dimers can be seen as a disquieting biomarker for different cancer types.

| Promoting effects of cancer on the development of VTE
Although the well-established correlation between cancer and the occurrence of VTE, underlying mechanisms that promote VTE are poorly understood. 4 Many previous studies have explored potential factors regarding the procoagulant impact of cancer on thrombogenesis in patients with miscellaneous types of cancer. 4-6 Current evidence suggests an inducible hypercoagulable status through the secretion of different factors or microparticles associated with endothelial function and platelet integrity, promoting angiogenesis, tumor invasion and growth. 6,51 In this context, the tumor-derived tissue factor-positive microparticles (TMPs) seem to play an important role in thrombosis. 6 Their primary role is believed to initiate thrombosis by providing negatively charged phospholipids, specifically phosphatidylserine, which facilitate the assembly and activation of enzymes involved in the coagulation cascade. While this association is most established in pancreatic cancer, TMPs are released by various cancer types, including brain, lung, ovarian and gastric tumors. 52 Their presence is linked to an elevated risk of VTE. 51 Moreover, the production of TMPs may increase with the severity of tumor grade and the potential for metastasis. 6 Furthermore, specific therapeutic measures, such as the application of immunomodulatory drugs or the initiation of chemotherapy, can promote thrombogenesis. 4,53 Recent investigations have found that some chemotherapeutic regimens like platinum-based therapies can interact with main pathways of coagulation, such as the activated protein C pathway. 6,54 The protein C pathway represents a major regulator of thrombin production that can be inhibited by damage to the endothelium due to chemotherapeutic treatment. 54,55 Other risk factors associated with venous or arterial thrombosis arise from vascular damage due to atherosclerosis, resulting in vascular disturbances with consecutive platelet adhesion. 56

| Diagnostic value of D-dimer testing in cancer patients
D-dimer testing represents a valuable indicator of the coagulation system with an established role in ruling out VTE. In cancer patients, D-dimers can also be used for accurate rule-out of VTE with a reported sensitivity of 100% (95% CI, 98%-100%) and a negative predictive value of 100% (95% CI, 97%-100%), if they are below the general rule-out cutoff of 0.5 mg/L or only slightly above (<0.6 mg/L). 36 In the case of highly elevated concentrations above the 10-times upper limit of normal (ULN), D-dimers are correlated with a high positive predictive value for patients being affected by DVT or PE. A recently published work revealed a ROC-optimal cut-off value of 9.9 mg/L at a positive and negative predictive value of 96% (95% CI, 85%-99%) and 78% (95% CI, 76%-80%), respectively. This threshold facilitated a diagnosis with specificities and positive predictive values of at least 90%. 36 However, it is important to acknowledge that the prevalence of elevated D-dimers is higher in cancer, which reduces the number of indivudals who meet the criteria for ruling out VTE based on a normal or slightly elevated D-dimer level. Research indicates that only 32% and 37% of patients with cancer and suspected VTE had D-dimer levels below 0.5 mg/L or below 0.6 mg/L, respectively. 36 On the contrary, the occurrence of D-dimers values exceeding 10-times ULN is relatively low. The optimal diagnostic strategy for patients with intermediate D-dimer concentrations is illusive. In such cases, clinical judgement is paramount to prevent excessive use of CT imaging and unnecessary anticoagulation. 36 Therefore, results of previous studies suggest that the practical use of D-dimer testing can be extended from its well-established role in ruling-out VTE to an aid for rule-in of cancer patients in the case of elevated D-dimer concentrations exceeding 10-times the ULN. In this study, approximately one-third of 526 cancer patients qualified for rule out suggesting that a relevant proportion of cancer patients could be successfully assessed in clinical practice. 36 A recent study suggests raising the cut-off for patients classified as 'PE unlikely'. This adjustment would lead to a moderate improvement in clinical usefulness and appears to be a promising and safe approach, particularly when considering age-dependent factors. 37 Further studies are necessary to improve the overall management of thrombotic risk in this population.

| Optimizing diagnostic precision: age-adjusted cut-off for D-dimer testing in cancer patients
Cancer is a strong risk factor for developing VTE and Ddimer levels are frequently elevated in cancer patients. Furthermore, the specificity of D-dimer testing decreases with age, potentially leading to unnecessary computed tomography angiography (CTA). 57 To improve the performance of D-dimer testing and increase its specificity, ESC Guidelines on PE recommend the application of age-dependent cut-offs for patients aged >50 years. This cut-off value for age-adjusted D-dimer is determined by multiplying the patient's age by 10 μg/L for individuals above 50 years of age. 57 Using the age-adjusted cut-off for D-dimer has been found to effectively increase the proportion of cancer patients in whom PE can safely be excluded without undergoing CTA. This may result in a reduction of diagnostic imaging, as well as a faster diagnostic workup. 57 It can avoid a significant number of inappropriate CTA scans at the cost of potentially missed small subsegmental PEs. 58 Further studies are required to determine its efficacy and to validate its widespread use in cancer patients.

| Prognostic potential of D-dimer testing as a predictor of adverse outcomes
Among cancer patients receiving outpatient chemotherapeutics, VTE is the second most frequent cause of death. 36,39 On the contrary, cancer patients who had a VTE are at increased risk for developing recurrent VTE and severe complications, such as critical bleeding. 1,13,37,59 A prospective population study showed that the mortality rate of cancer patients 1 year following VTE is 2.5-times higher compared to all cancer patients. 60 However, elevated D-dimers in general correlate with a poorer prognosis. During a median follow-up of 30 months, a significant positive correlation has been reported between D-dimers and the reoccurrence of VTE (p = .0299) as well as all-cause mortality in cancer patients, regardless of whether they had VTE or not (p < .0001 and p = .0008, respectively). 36 So far, it is unclear whether D-dimers only indicate a more aggressive cancer type or a higher prevalence of thromboembolic complications. 36 In this context, VTE seems to be a hallmark of an advanced cancer stage. 60 D-dimer levels can be seen as a predictor of metastatic status and cancer progression. 12,14,61 Previous studies revealed a difference in D-dimer plasma levels based on metastasis and TNM stage among cancer patients. 14 They reported a significant association between increased plasma D-dimer levels and TNM stage in colorectal cancer patients. 62 D-dimers can also be used as a predictive marker of clinical stage among gastric and breast cancer patients. 45,63 Elevated D-dimer levels were able to predict positive lymph node involvement and the F I G U R E 4 Proposed clinical decision rule for deep venous thrombosis (DVT) and pulmonary embolism (PE) in cancer patients (customized approach). Two alternative classification schemes can be used for clinical probability assessment, that is a three-level scheme (clinical probability defined as low, intermediate, or high) or a two-level scheme (PE unlikely or PE likely). The established age-independent rule-out cut-off level is 0.5 mg/L. 1 In the event of a negative CTPA in patients with a high clinical probability, it may be advisable to consider additional investigations such as tests for DVT. CTPA, computed tomography pulmonary angiography; DVT, deep vein thrombosis; PE, pulmonary embolism; ROC, receiver operating characteristic.
presence of lymphovascular invasion among breast cancer patients. 45,64 Even in paediatrics, D-dimer levels can be seen as a predictor of cancer stage. Higher D-dimer levels were observed in paediatric patients with hepatoblastoma stage 4 patients compared with stage 1-3 patients. 65 A recent study could display that high levels of circulating D-dimers increase the mortality rate in patients with lung cancer. 47,66 These results indicate that the D-dimer level may have the potential to predict the probability of metastasis, progression and mortality in different types of cancer. 14,67,68

| Venous tromboprophylaxis in cancer patients
Recommendations for venous thromboprophylaxis in cancer patients are typically based on individual patient characteristics and risk assessment. The complete management of VTE in cancer patients involves two important aspects. First, the identification of these patients who would benefit the most from pharmacological prophylaxis. Second, the implementation of effective treatment strategies to minimize the risk of VTE recurrence and mortality in these patients. 69 The American Society of Clinical Oncology (ASCO) guidelines recommend offering pharmacological thromboprophylaxis to hospitalized cancer patients with additional risk factors such as limited mobility, as well as those without additional risk factors, provided there are no bleeding or other contraindications. 69 On the contrary, pharmacological thromboprophylaxis should not be routinely provided to all cancer outpatients. 69

| Perspectives
The main hurdle in the diagnostic process for cancer patients with suspected PE is the timely differentiation between individuals at high risk of the disease and those without it. 37 The use of the Wells pretest probability model, along with a negative D-dimer test, has shown reduced clinical utility in cancer patients compared to those without cancer. 35,36 To address this issue, one possible solution is to develop a modified cancer-specific clinical probability score or to adjust D-dimer thresholds. This could help improve the accuracy and effectiveness of the diagnostic process in this particular population by incorporating cancer-related factors into the assessment. 2,37 Further research and development in this area are crucial to investigate new diagnostic approaches and to establish a more reliable and personalized diagnostic approach for cancer patients with suspected VTE. In addition, further research is necessary to explore the potential underlying mechanisms responsible for the poorer prognosis of cancer patients with VTE and the role of D-dimers as a predictor of cancer stage. 14,36,59,60,70

| Conclusions
In summary, D-dimer testing and the Wells score are valuable tools in the exclusion of VTE. However, their diagnostic efficacy is diminished in cancer patients compared to individuals without cancer, leading to reduced clinical utility. Nonetheless, D-dimer levels exceeding 10-times the ULN indicate a high probability of DVT or PE and should be taken seriously in cancer patients. D-dimer levels have significance beyond their diagnostic role in ruling out VTE. They can also serve as predictors of cancer stage and act as prognostic markers. Efficient identification of cancer patients at a significant risk of developing VTE requires further exploration of novel diagnostic strategies and prediction rules in future studies. It is important to continue investigating and refining diagnostic approaches in order to enhance the management and outcomes for cancer patients at risk of VTE.

AUTHOR CONTRIBUTIONS
All authors have contributed significantly to this work, participating in the conception, design, analysis, interpretation of data and writing/editing of this manuscript. J.G. and V.K. wrote the manuscript and designed the research.

PERMISSION TO REPRODUCE MATERIAL FROM OTHER SOURCES
Material from other sources has not been used.