Survival rates for children with cancer have improved over the past decades [1,2]. Recent trends show that although survival rates for children with primarily leukemias and lymphomas have improved, 5-year survival rates for pediatric solid tumors have not changed over the past 2 decades. The limited numbers of clinical trials on pediatric solid malignancies as well as difficulty in discriminating and diagnosing solid masses by using standard techniques have hindered progress in this area. Therefore, it is imperative to not only accurately determine the type of masses but also detect tumor recurrences.

Serum tumor markers offer great promise in the management of children with solid malignancies and are integral to disease diagnosis and prediction of treatment response. These cancer biomarkers are biologically measured substances that are expressed by malignant tissues, circulating tumor components, or generated by the host in response to the tumor, and constitutively serve as essential tools to aid clinicians in making diagnosis, staging, and risk assessments. The nonspecific property of some markers to be differentially expressed in other tissues limits their clinical use and hampers accurate diagnosis of disease information. Although there is a wide spectrum of pediatric solid tumors, this review discusses only current trends in the use of serum biomarkers for detection and evaluation of solid tumors in the chest and abdomen of children.

Mediastinal tumors are relatively common in infants and children and represent a heterogeneous group of asymptomatic or potentially life-threatening neoplastic lesions that may carry complex diagnostic and therapeutic dilemmas. In general, imaging with a chest radiograph (anterior-posterior and lateral projections) and a computed tomography (CT) scan gives anatomic guidance for primary tumor location and information on mass morphology, extent, and relation to surrounding structures. Depending on tumor location within 3 different mediastinal compartments (anterior, middle, and posterior), a differential diagnosis may be made along with information on age and clinical history. Anterior mediastinal tumors include germ cell tumors (GCTs), thymic-related, cystic hygromas, and lymphomas. Common masses in the middle mediastinum are lymphomas, bronchogenic cysts, and granulomatous disease. Posterior mediastinal masses are mainly neurogenic tumors and represent the most common tumors in the mediastinum [3].

Pediatric solid abdominal neoplasms encompass a wide spectrum of diseases ranging from lesions causing significant morbidity and mortality [advanced stage neuroblastoma (NB)] to conditions that can be corrected by surgery alone [favorable-stage Wilms tumor (WT)]. The physician needs to determine the nature of the mass in a timely, safe, and cost-effective manner. Likely abdominal malignancies diagnosed in children include NB, WT, hepatoblastoma (HB), lymphoma, soft tissue sarcomas, and GCTs. As for thoracic masses, diagnostic imaging plays a major role in the diagnosis and management of pediatric abdominal masses, local disease staging, identification of distant metastases, and monitoring response to therapy. However, imaging techniques cannot always accurately distinguish between benign and malignant masses and may be discrepant from histologic staging [4]. Table 1 summarizes these biomarkers. In the following sections, we describe serum markers available for pediatric solid tumors in the chest and abdomen.

Serum biomarkers are useful tools to differentiate solid tumors in children. There is an abundance of literature on serum biomarkers, but this review has focused on a select handful of clinically relevant markers used for the detection, surveillance, and survival prediction of common pediatric malignancies of the chest and abdomen. Although these laboratory tests are helpful in discriminating masses, it must be recognized that some markers do not always accurately indicate disease, but are in general reliable when used in combination with other diagnostic tools. They can be used to obtain information required for planning future cancer treatment.

Nevertheless, professional societies have formal practice guidelines for the use of most tumor markers based on evidence-justified practice, which culminates pertinent, trustworthy information by systematically acquiring, analyzing, and transferring research findings into clinical, management, and policy arenas [144]. For instance, a wealth of publications on GCTs serum markers supports the use of hCG, AFP, and LDH serum concentrations for clinical management. Recent reviews by the American Society of Clinical Oncology and the National Academy of Clinical Biochemistry have outlined updated practice guidelines in adults [145,146]. These studies underscore the relevant issue of comparing age-related variation in laboratory values. Although most studies on existing GCTs tumor markers in children have similar observations, there are considerable differences in tumor biology in children and the extrapolation of results from studies in adults to children is controversial. For example, 40%–55% of GCTs in children are found at extragonadal locations as against the 5%–10% of GCTs found in adults at the same location [147–149]. This difference is likely due to variations in the maturity of the germ cells that give rise to the tumors in these age groups [150]. GCTs in children likely originate from a primordial germ cell (PGC) that underwent immediate reprogramming to become a pluripotent embryonic germ cell, whereas GCTs in adolescents and young adults most likely originate from more mature PGCs, which may be unable to survive outside of the normal niches of the ovary and testis or specialized sites such as the thymus in the case of mediastinal GCTs [151]. Additionally, the absence of BRAF mutations in pediatric GCTs in comparison to adult tumors was recently reported and further supports evidence for patient age-related biological differences [152]. This translates to distinct clinical/genetic profiles for pediatric and adult GCTs, and, especially in cases of AFP wherein these serum levels are developmentally regulated, highlights the significant role of the age of patients in determining normal baseline serum levels of some tumor markers.

Although age-adapted reference intervals are a prerequisite for proper interpretation of laboratory data, other sources of laboratory testing variation include preanalytical, total random analytical error, and intraindividual normal biology. The discrimination of different molecular forms of tumor markers may also have ramifications at the clinical level. For example, some question the validity of differences in hCG concentrations, as the recognition of hCG variants [nonnicked and nicked intact hCG (hCG and hCGn), nonnicked and nicked free β-subunit (hCGβ and hCGnβ), and regular and hyperglycosylated (or large) free hCGα] can cause discordance in hCG immunoassay results. Cao and Rej identified the following factors that contribute to the incorrect reporting of hCG results by laboratories: (a) complexity of the hCG molecule and confusion of nomenclature among various forms of hCG; (b) laboratory personnel’s lack of awareness of the distinctions in the forms of hCG and failure to recognize the specificity of assays for measurement of different forms; (c) lack of clarity and uniformity in manufacturers’ reagent labeling; and (d) lack of information on the specificity of each method for the various forms of hCG in most product inserts [153].

It is vital to accurately detect the specific hCG-related molecule in patients with malignant disease. Immunochemical characterization has shown that though intact hCG is abundantly produced by the placenta and germ cell tumors, it is the serum free β-hCG subunit (independent of the common glycoprotein hormone α subunit [GPHα]) that is also a tumor marker in many non-trophoblastic tumors, which include common epithelial cancers. Therefore, it is critical to know which hCG isoforms are measured in order to avoid false-positive results, as most automated commercial laboratory tests, point-of-care tests, and over-the-counter tests are limited in what they detect, as they focus only on regular hCG.

Despite considerable potential limitations to current pediatric tumor markers, continued progress in tumor marker discovery will likely come from rigorous translational investigations that will integrate multidimensional analyses from various emerging technologies to deliver more specific predictive and prognostic markers. Undoubtedly, next-generation biomarkers will likely stem from the growing omics technologies, which will amalgamate large-scale genomics, transcriptomics, and proteomics to promote personalized cancer care. Consequently, proteomics based biomarker discovery is one of the omics platforms that has garnered much interest in many diseases, including cancer. The enthusiasm towards cancer-related proteomics applications is due to the role proteomics can provide in bridging the gap between genetic alterations underlying cancer and cellular physiology, such that a cell’s genome is relatively stable and gives information of the organism’s gene expression while the proteome is dynamic, complex and variable. Moreover, there is a growing opinion in favor of the role of targeting proteomic measurements for improving blood-based human diagnostics [154,155]. Blood has been the logical biomarker source, as blood equilibrates with tissues and generally harbors proteins that may be identified to yield a specific proteome pattern related to a distinct pathologic process occurring in the body. To date, blood-based proteomic biomarker efforts have had little success, in large part because the relatively small number of highly abundant proteins make the reliable detection of low abundant disease-related proteins challenging due to the wide dynamic range of concentrations of proteins in a blood sample. Given the substantial complexity of plasma and the vast dynamic range of protein abundance, approaches used to simplify and increase the depth of proteomic analysis include fractionation strategies, targeted protein subpopulation enrichment (phosphoproteins, glycoproteins, etc.), and differential quantifying protein methods such as isobaric tags for relative and absolute quantification (iTRAQ).

In addition, other novel work in biomarker discovery involves proximal fluid proteomics. Proximal fluid, the fluid derived from the extracellular milieu of tissues, contains a large repertoire of shed and secreted proteins that are likely to be present at higher concentrations than in the plasma or serum. It has been hypothesized that many, if not all, proximal fluid proteins exchange with peripheral circulation, which has triggered the use of proximal fluids as a primary sample source for protein biomarker discovery [156].

Lastly, exciting opportunities in nanoscale exploration open new strategies for serum protein analysis. For example, Pujia and colleagues recently described a tool based on biodegradable nanoporous nanoparticles (NPNPs) that allows the harvesting of low-molecular-weight fractions of crude human serum or plasma [157]. NPNPs with a diameter of 200 nm and pore size of a few nanometers have been obtained by ultrasonication of nanoporous silicon. When incubated with a solution, NPNPs harvest only molecules that are small enough to be absorbed into the nanopores. These molecules can then be recovered by centrifugation and dissolved in water to give samples for further analyses. The development and use of novel methods in serum biomarker discovery may include strategies that use nanostructured materials and other high-throughput methods such as protein arrays, multiplexed protein assays, and chip-based proteomic platforms. For instance, utilizing protein-chip based array proteomic technologies, Wang et al. [158] performed protein profiling of serum samples from pre-surgery and post-surgery patients with WT and healthy controls. Two peaks (11,526 and 4,756 Da) showing significant differential expression not only between pre-surgery and control sera but also between pre-surgery and post-surgery sera were identified and characterized as serum amyloid A1 (SAA1) and apolipoprotein C-III (APO C-III).

In the continued search for ideal tumor markers in children, it is essential to pursue study designs that are appropriately powered and lend themselves to independent replication. There are several criteria for any biomarker to be considered clinically relevant. Sturgeon and colleagues recently updated the National Academy of Clinical Biochemistry laboratory medicine practice guidelines and quality requirements for the use of tumor markers. The most important feature with regard to these biomarkers is the ability to affect patient management [146]. Savage and Everett have defined the other challenges that are faced in biomarker research in children [159]. In particular, an important issue with regard to proteomics is the validation of promising candidates for pediatric diseases by blood-based profiling [160]. Proteomic profiling on children has been criticized as limited sample sizes and volumes are available for analysis, which may reduce the accuracy of the validation. Also, markers in children are discovered in the context of growth and development; few studies have reported serum or plasma proteins from healthy children. It is important to obtain serum or plasma profiles to determine age-related changes in order to better comprehend whether candidate proteins are related to disease or to normal growth. Currently, multiple profiling projects on the human plasma proteome are underway [161–163], and recent study has reported a web-based database of human plasma proteins from normal individuals [164]. Such a resource specific to children needs to be established to enable proper comparisons to interpret blood samples from healthy children and cancer patients in order to drive the identification and validation of biomarker targets.

The progress toward early diagnosis, curative therapy, and favorable outcomes in childhood solid cancer patients and survivors depends not only on developing continued strategies to identify biomarkers for more precise diagnosis, targeted therapy, and possibly prevention, but also designating additional funding or redistributing existing resources for furthering the study of childhood malignancies that continue to have static or poor survival trends. For example, 5-year survival rates for pediatric and adolescent sarcomas (Ewing, osteosarcoma, and rhabdomyosarcoma) have not drastically changed over the past 2 decades as compared with those for other solid tumors such as non–central nervous system GCTs and Wilms tumor [2]. Given the paucity of data on biomarkers and the challenging clinical management of children with metastatic sarcomas, a cross-disciplinary biological approach that comprehensively assesses DNA, RNA, protein, and metabolites to identify molecular drivers and markers is warranted to conduct small, short, and more economical individualized clinical trials. We foresee molecular data on solid pediatric cancers being translated into practical diagnostic and treatment modalities in order to optimize the success rate and hasten the implementation of effective therapies into the clinic.