PCNSTs represent a complex heterogeneous group of pathological entities that may be benign, malignant or of unpredictable evolution [1–10]. These tumors represent a major public health problem [11]. The term “primary brain tumor” has been defined in numerous ways in the literature. The primary difficulty in building a tumor registry and/or a data base is to define the type of tumor to be recorded.

Recent publications [12–20], the classification system of the World Health Organization [1, 2, 6], and the European recommendations for coding tumors of the brain and CNS [21] include all primary benign and malignant tumors located in the CNS, including its envelopes and the beginning of the nerves localized in the skull and spine. PCNSTs include tumors of neuroepithelial tissue (gliomas and all other neuroepithelial tumors), tumors of the meninges (meningiomas, mesenchymal tumors and other tumors of the meninges), tumors of the cranial and paraspinal nerves, lymphomas and hematopoietic neoplasms, and others.

By definition, this excludes metastatic tumors. But even now, there are still discrepancies concerning the inclusion of lymphoma, pituitary and pineal glands, and olfactory tumors of the nasal cavity, in the different databases and registries. It is important to note that some institutions account for primary malignant tumors, only.

Coding systems have been introduced as an indispensable interface between pathologists and cancer registries. The international classification of diseases for oncology (ICD-O) was established more than 30 years ago. It assures that histopathologically stratified population-based incidence and mortality data become available for epidemiological and oncological studies. The histology (morphology) code is increasingly complemented by genetic characterization of human neoplasms. The ICD-O histology codes have been adopted by the systematized nomenclature of medicine (SNOMED), issued by the College of American Pathologists. ICD-0-3 and SNOMED codes are available in Louis et al. [1, 2] and in Rigau et al. [19]. All glioma morphology codes are shown here (see Table 2.3) [23].

The ICD-O topography codes largely correspond to those of the tenth edition of the International statistical classification of diseases, injuries and causes of death (ICD-10) of the WHO. ICD-O-3 topography codes include: brain (C71.0–C71.9), meninges (C70.0–C70.9), spinal cord, cauda equina, cranial nerves, and other parts of the central nervous system (C72.0–C72.9) [24]. Now, many registries use these codes, but some registries still record malignant tumors only [25]. Some registries (e.g., Central Brain Tumor Registry of the United States –CBTRUS-) include more topography codes as pituitary and pineal glands (C75.1–C75.3), and olfactory tumors of the nasal cavity [C30.0 (9522–9523)].

Another challenge for registries and databases is to record and to detail all cases of defined tumors (i.e., all histological types and all histological subtypes, when it is possible).

We present the proportion of each major type of PCNST and the distribution of DLGG in the French Brain Tumor DataBase (FBTDB) for the 2006–2011 period (Fig. 2.1a–c). We point out (1) FBTDB records cases with histological confirmation only, knowing that usually, registries record the cases with and without histological confirmation, and (2) the distribution of the different subtypes of PCNST is similar in France and in the USA, except for oligodendroglial tumors (see [19]). Recently, most studies (e.g., [26, 27]) have reported a recent increase of oligodendroglial tumors in comparison to astrocytic tumors; however, French neuropathologists are more influenced by the classification proposed by Daumas-Duport et al. [28] than American neuropathologists.

We could see here the difficulties to compare data from different institutions when histological diagnosis is not completely reproducible [29].

An additional difficulty to know the overall proportion of DLGG is that (1) some institutions use terminology such as “low grade gliomas” (LGG, see below) and (2) when some information are not detailed enough, registries include some cases in astrocytoma not otherwise specified (NOS) or in glioma NOS, even in unspecified neoplasm.

However, we can consider that the overall DLGG represents about 13–16% of all gliomas.

The number of males and females and the median age at diagnosis (MAD) for DLGG in FBTDB (period: 2004–2009) [23] are shown in Table 2.3. Sex ratios, MAD in CBTRUS [15, 17] and in FBTDB are shown in Table 2.4. Sex ratio (male/female) for all DLGG is very similar in CBTRUS (1.33) and in FBTDB (1.32). It is the same for MAD for each sub type of DLGG. One can hypothesis that these similarities across the Atlantic could argue for genetic components of these tumors. To be rigorous about classification system, we can notice (1) gemistocytic astrocytoma were not included in DLGG in 2009 and 2012 CBTRUS reports (probably, because gemistocytic astrocytomas are more aggressive tumors, CBTRUS counted gemistocytic astrocytomas with anaplastic astrocytomas [note: 2015 CBTRUS report includes gemistocytic astrocytomas in diffuse astrocytomas, as in the 2007 WHO Classification] (2) CBTRUS does not separate oligoastrocytomas from anaplastic oligoastrocytomas, because these tumors have the same SNOMED code (9382/3) (see Tables 2.1 and 2.3).

Another important thing to notice when one speaks of MAD is how diagnosis is defined. When surgery is performed, registries and the large majority of databases use usually the date of the histological diagnosis. If we refer to clinical studies, some of them specify the age of radiological diagnosis, or sometimes, the age at first symptom. For example, the “Reseau d’étude des gliomes” (REG, a French consortium that studies DLGG) considered several starting points for their analyses: the date of available radiological diagnosis (which did not differ significantly from the date of clinical onset) and the date of first treatment. Among 1091 patients (collected at adult age), it was noted that median age (MA) at discovery was 37 years old (rage: 4–75) and the MA at first treatment was 44 years old (range: 18–76) [30]. Probably, this discrepancy would be decreased in the future, because clinicians seem to treat DLGG earlier than in the past.

Surgical data concerning DLGG are rare in population studies. Table 2.3 shows the percentages of resection versus biopsy for 14,083 gliomas and all subtypes (included DLGG) (FBTDB, collected data from 2004 to 2009). Of the 54 institutions that participated, the proportion of resection versus biopsy varied considerably from one institution to another (data not shown). We can also note that at least 27% of all glial tumors were cryopreserved (see [19, 23]). This is very important for future biological studies.

For many years, epidemiological studies have investigated all gliomas or even all PCNST. More recently, the term “low grade glioma (LGG)” has been introduced. LGG is probably used as a practical definition being quite simply in opposition to the term “high grade glioma (HGG)”. LGGs are slow growing, intrinsic lesions that contain glioma cells. Referring to the WHO classification of tumors of the CNS by Louis et al. [1], LGG may be defined by grade I (GI) and grade II (GII) gliomas and includes subependymal giant cell astrocytoma (GI), pilocytic astrocytoma (GI), pilomyxoid astrocytoma (GII), diffuse astrocytoma (GII), pleomorphic xanthoastrocytoma (GII), oligodendroglioma (GII), oligoastrocytoma (GII), subependymoma (GI), mixopapillary ependymoma (GI) and ependymoma (GII). Some authors include gangliomas, desmoplastic gangliomas [31], and even dysembryoplastic neuroepithelial tumors (DNETs) [32]. However, many studies on LGG exclude ependymomas and refer to astrocytic and/or oligodendrocytic tumors only. Furthermore, some pediatric studies mainly focus on pilocytic astrocytic tumors, and some adult studies often focus on GII astrocytic and/or oligodendrocytic tumors only.

Here, the term “diffuse low grade gliomas (DLGG)” includes grade II gliomas for diffuse astrocytomas, oligodendrogliomas and oligoastocytomas (mixed gliomas). This term is much more precise—because in essence it excludes well delineated grade I and II gliomas, and ependymomas with a different natural history. One benefit of this term is also to ignore the difficulties to differentiate an oligoastrocytoma from an astrocytoma or an oligodendroglioma. Moreover, even if heterogeneities are present in the GII glioma group of diffuse astrocytomas, oligodendrogliomas and oligoastocytomas, these tumors are often studied together. Usually they concern middle-aged adults with professional activities. The main clinical presentation is epilepsy with or without mild cognitive disorders. Focal deficit and/or raised intracranial pressure are possible but very infrequent. And the majority of these tumors have typical imaging characteristics on MRI as non-enhancing infiltrative lesions involving the white matter and frequently extending to the cortical surface [33]. Moreover, now the new WHO CNS Classification specifies the group of diffuse astrocytic and oligodendroglial tumors [2], Tables 2.1 and 2.2.

As mentioned previously, incidence data could vary from countries because the coding rules are not exactly the same all over the world. For example, referring to the 2007 WHO Classification, diffuse astrocytoma include fibrillary astrocytoma, gemistocytic astrocytoma, and protoplasmic astrocytoma, but some registries include astrocytoma NOS also. Classically, registries record cases with and without histological confirmation. It means that a number of cases of tumor are diagnosed by radiology, or even by clinical data only, or even by death certificate only. This is why the author decides to show data from (1) registries (e.g., Central Brain Tumor Registry of the United States—CBTRUS, and Austrian Brain Tumor Registry—ABTR) and (2) from histological population-based study (e.g., French Brain Tumor DataBase—FBTDB) that include all cases with histological validation only (Table 2.5). Histological population-based studies are more accurate but underestimate the true incidence [34–36]. Some other publications (e, g., [20, 37, 38]) present incidence data on DLGG, and one recent publication summarizes incidence data of glioma and some sub-types in different countries [39].

When we want to compare the incidence of one specific tumor from one country to another, another issue to consider is the differences in the two populations. Crude rates, whether they represent incidence, mortality, morbidity or other health events, are summary measures of the experience of populations that facilitate this comparative analysis. However, the comparison of crude rates can sometimes be inadequate, particularly when the population structures are not comparable for factors such as age, sex or socioeconomic level. For example, median age at diagnosis of glioblastoma is 64 years; because the age distribution is different in occidental word than in developing countries, we can not compare the crude rate of glioblastoma in US to the crude rate of glioblastoma in India. To enable international comparison, age-standardized rates from the FBTDB data were calculated with the USA, Worldwide, and Europe populations as references (Table 2.5).

Moreover, from epidemiological point of view, the difficulty in grading the diffuse gliomas in grade II vs. III by pathologists in some cases could modify the incidence of DLGG vs. DAG (e.g., microfoci in DLGG [40]). In the past, neurosurgeons gave only small samples to pathologists; now bigger samples of the tumor are analyzed by the pathologist, and pathologists have more sophisticated techniques. So we could imagine that incidence of DLGG would decrease slightly while incidence of DAG would increase slightly.

The 2016 WHO Classification integrates biology but the difficulties still persist in differentiating GII and GIII in some cases.

In occidental world, the incidence rate for all DLGG is between 0.6 and 1.3/100,000 person-years. So, we can hypothesize that the value of 1/100,000 person-years, or a little bit less, could be considered as a good approximation.

So, this first conclusion about incidence of about 1/100,000 person-years for DLGG in occidental world implies few comments and questions:

The variation of incidence between each subtype of DLGG seems more important than the variation of the overall incidence. It means that histological criteria between astrocytoma, oligodendroglioma and oligoastrocytoma are probably variable from one country (or one center) to another. Does the new WHO Classification will do better? We cannot answer yet.

What can explain the (relatively small) discrepancies between the overall incidence of DLGG from one country (or one area) to another, in the occidental world? Only methodology? Or real difference? We cannot answer yet. But we can notice that incidence of overall DLGG is higher (1) in men than in women (see sex ratio of M/F of about 1.3), and (2) in White than in Black in USA. CBTRUS data [15] give incidence rates (per 100,000 person-years, age-adjusted to the 2000 U.S. standard population): protoplasmic and fibrillary astrocytoma: 0.10, oligodendroglioma: 0.31, mixed glioma: 0.19; Male/Female: 0.13/0.08, 0.34/0.27, 0.24/0.16; White/Black: 0.12/0.04, 0.34/0.13, 0.21/0.08 for the same histology respectively. Furthermore, the FBTDB showed that the geographical distribution by region of the DLGG had significant differences, with higher incidence rates in Northeast and central parts of France [34].

From epidemiological point of view, two additional points need to be mentioned. First, the incidental findings on MRI are increasing with the use of MRI for many other clinical situations (brain traumas, headaches, neurologic disorders, etc.) and even for research purposes [41–43]. Secondly, DLGG patients have a long silent stage before to be symptomatic, and some teams discuss screening and preventive therapeutic management [44–46]. So, we can imagine that the incidence of DLGG could rise in the future.

In 2016, many interesting publications are available in the literature about gliomas survival. But unfortunately, most of them include a lot of different subtypes. Many publications analyze only clinical, or therapeutic, or biologic prognostic factors. So, it is very difficult to conclude precisely. Of course survival of DLGG is better than diffuse high grade gliomas, but DLGG is still an ultimately fatal disease.

US survival rates for DLGG [diffuse astrocytoma (including fibrillary, gemistocytic, protoplasmic, and NOS astrocytomas), oligodendroglioma, and mixed glioma (including oligoastrocytoma grade II and III because they have the same SNOMED code) and selected glioma subtypes [i.e., anaplastic oligodendroglioma, anaplastic astrocytoma, glioblastoma, and glioma malignant NOS] for comparison, are presented in Table 2.6 (adapted from 2015 CBTRUS report [18], period 1995–2012). The estimated 5-/10-year relative survival rates for diffuse astrocytomas and oligodendrogliomas are 47.9/37.6 and 79.8/64.0%, respectively. Three points could be noted: (1) the estimated 5- and 10-year relative survival rates for mixed gliomas (included oligostrocytomas and anaplastic oligoastrocytomas) are 62.0/47.8%, respectively; (2) the number of glioma NOS is particularly high and (3) survival of glioma NOS is not so poor as glioblastoma survival.

In the population-based study of the Canton of Zurich, Switzerland (N = 122 DLGG, period 1980–1994), the survival rate (mean follow-up 7.5 ± 4.8 years) was highest for patients with oligodendroglioma (78% at 5 years, 51% at 10 years), followed by those with oligoastrocytoma (70% at 5 years, 49% at 10 years) and fibrillary astrocytoma (65% at 5 years, 31% at 10 years). Survival of patients with gemistocytic astrocytoma was poor, with survival rates of 16% at 5 years and 0% at 10 years [38].

Elsewhere in Europe, data from cancer registries are not specific for DLGG and vary with periods of time and regions. For example, the Danish registry, one of the oldest registries, compared the overall survival of patients with oligodendroglial tumors (oligodendrogliomas and anaplastic oligodendrogliomas) during the periods 1943–1977 and 1978–2002. The median survival increased from 1.4 years (95% confidence interval [CI], 1.0–1.6) to 3.4 years (95% CI, 2.6–4.2) during the period of study [47]. More recently, EUROCARE group (included data from 39 cancer registries located in different regions of Europe) showed that estimates of 5-year relative survival rates (95% CI) for patients with oligodendrogliomas/anaplastic oligodendrogliomas, alive in 2000–2002, were 74.1 (64.4–81.8)/35.1 (21.2–49.5) in Northern Europe, 65.8 (57.5–73.0)/35.5 (24.4–46.9) in UK and Ireland, 75.5 (61.8–85.2)/29.7 (13.4–48.3) in Central Europe, 47.8 (32.4–62.0)/6.1 (1.3–16.6) in Eastern Europe, 63.8 (51.4–74.1)/33.3 (14.7–53.6) in Southern Europe, and 67.2 (62.5–71.6)/31.5 (25.0–38.3) in all cases [48]. The last data of primary malignant brain tumors from EURCARE (including 58 cancer registries, period 2000–2007) have grouped many glioma types, and it is not possible to separate DLGG. But for the group of “oligodendroglioma, anaplastic oligodendroglioma, oligoastrocytoma and anaplastic oligoastrocytoma”, the 5-years relative survival for the subgroup of 45–54 years old patients and the subgroup of 55–64 years old patients were 51.4 and 31.0% in the period 1999–2001, versus 60.4 and 38.5% in the period 2005–2007 with p = 0.005 and p = 0.02, respectively [49]. In the work by Crocetti et al. [50], data from 76 cancer registries out of the 89 that accepted to participate in the RARECARE project were considered. The estimated 5-year relative survival was 14.5% for astrocytic tumors (42.6% for low grade astrocytomas, 4.9% for high grade astrocytomas, and 17.5% for gliomas NOS), and 54.5% for oligodendroglial tumors (64.9% for low grade and 29.6% high grade). Survival data for gliomas including data for some subtypes of DLGG are available in few other countries (e.g., England, Korea, Netherland, and Sweden [51–54]). Ostrom et al. [39] compared some of them.

US population data show that all DLGGs have survival averaging approximately 6 [54]–7 years [55], although variation in survival is quite large with at least 20% of patients surviving for two decades [55]. The median survival for patients with astrocytoma, mixed glioma, and oligodendroglioma is 5.2, 5.6, and 7.2 years, respectively.

One important question is to know if survival of DLGG is improved with the new medical technology. The answer of this question is not so easy.

Data from the Surveillance, Epidemiology and End Results (SEER) program of the National Cancer Institute suggest that for the majority of low grade glioma patients, overall survival has not significantly improved over the past three decades [56]. Data from other, but smaller works showed a modest improvement at least for tumors with oligodendroglial component [47, 49, 53], and some important clinical studies showed a median survival longer than 10 years (e.g., [30]).

Generally speaking, the length of the mean survival of DLGG, the heterogeneous medico-surgical care, the choice of the best personalized therapeutic, the acquisition of experience of new technologies (awake surgery, second or third surgery eventually, chemotherapies, new modalities of radiotherapy, etc.), needs a follow-up of at least 10–20 years to demonstrate an improvement at population level. Moreover, very few registries collect many prognostic factors as functional status, biology, quality of resection, all therapeutic lines, etc. to compare survival.

Prevalence rates are ideally suited to provide an overall estimate of cancer survivorship and direction for health planning as they reflect the complex relationships between incidence, survival, and population demographics—and hence to provide valuable information to the research and medical community. But prevalence data for PCNSTs are limited and very difficult to obtain. In theory, this would imply that the registration of cases is (and has been) exhaustive for many years (to account for long survivors), and that the histological classification systems have not changed over this long period. In 2001, Davis et al. showed that the prevalence rate for all PCNST was 130.8 per 100,000 with approximately 350,000 individuals estimated to be living with this diagnosis in the United States in 2000. The prevalence rate for primary malignant tumors was 29.5 per 100,000, the prevalence rate for primary benign tumors was 97.5 per 100,000, and 3.8 per 100,000 for primary borderline tumors [57]. The same group published new prevalence data in 2010. On the basis of the sum of non-malignant and averaged malignant estimates, the overall prevalence rate of individuals with a PCNST (as defined by CBTRUS) was estimated to be 209.0 per 100,000 in 2004 and 221.8 per 100,000 in 2010. The female prevalence rate (264.8 per 100,000) was higher than in males (158.7 per 100,000). The average prevalence rate for malignant tumors (42.5 per 100,000) was lower than for non-malignant tumors (166.5 per 100,000) [58]. In Europe, Crocetti et al. [50] published that the estimated prevalence rate for all astrocytic tumors of CNS was 20.4/100,000, and the estimated prevalence rate for oligodendroglial tumors of CNS was 2.7/100,000, with an incidence rate (per 100,000 person-years, age standardized on European population) of 4.4 and 0.4, respectively.

If until now, no specific prevalence data for DLGG are available in large population, we could use a crude approximation. If the average duration of disease and the population of patients are stationary, the prevalence can be estimated by the incidence (I) and the mean duration of disease (D) with the following equation: P ≈ I×D. Given the survival data (see previous and following paragraphs and without taking into account that survival could be improved), a crude approximation of the DLGG mean duration could be about 7–12 years. An estimation of the incidence rate for DLGG was ≈1/100,000 person-years (see above). So according to this crude approximation (and with I ≈ 1/100,000 person-years), the approximate value of the prevalence rate for DLGG would be about 10/100,000, or even more if we consider that the mean survival is longer. It is important to notice that the prevalence rate for DLGG is higher than the prevalence rate for glioblastoma.

On the other hand, as it has been noted by Mandonnet et al. [59], DLGG is a progressive primary brain tumor for which several stages can be discerned (i.e., one of them is a long clinical silent stage). Now, it is possible to have an estimation of the prevalence of silent DLGG by MRI performed in healthy subjects. In the conditions of the experiments, the prevalence of DLGG was between 0.1 and 0.2% of the studied population [41–43].

Age is one of the most important spontaneous prognostic factors for DLGG. Survival of diffuse astrocytomas and oligodendroglioma by age groups (CBTRUS 2015 data, period 1995–2012, [18]) are shown in Table 2.7. In the study by Claus and Black [55] entitled “survival rates and patterns of care for patients diagnosed with supratentorial low-grade gliomas (Data from the SEER Program, 1973–2001)”, improved survival was significantly associated with female gender (hazard ratio [HR], 0.84; 95% CI, 0.74–0.95), younger age, white race (HR, 0.70; 95% CI, 0.54–0.93), histology, and later year of diagnosis. In Europe, Crocetti et al. [50] noted that for glial tumors, the 5-year survival was slightly higher for women (20.7%; 95% CI 19.6–21.9) than for men (18.7%; 95% CI 17.8–19.7). Sant et al. [48] found also slightly better survival for women than men only for malignant PCNSTs. For many cancers, women survive longer than men, and this has been attributed to lower prevalence of comorbidities in women, a better performance status (allowing full application of effective surgical and adjuvant treatments) as well as a better “resistance” to disease [60].

The clinical and neurological status, before and/or after an oncological treatment, classically influences survival [61, 62]. The presence of a neurological deficit increases with age, tumor extension and mass effect [63]. At time of diagnosis, the existence of epilepsy is inversely linked to the presence of a deficit, and consequently carries a favorable prognostic value when isolated [64–66].

DLGGs are commonly located in or close to eloquent areas, i.e. those areas of the brain involved in motor, language, visuospatial and cognitive functions [30, 62, 67–69]. Larger tumors and tumors crossing the midline correlate with a shorter survival [64]. Growth rates are inversely correlated with survival [70–72]. It is important to note that DLGGs show a constant linear growth before malignant transformation. Very slow progression is possible, but these tumors always grow. The average slope is about 4 mm of mean diameter per year before and/or after surgical resection (without adjuvant therapy) [73–75].

In a recursive partitioning analysis, Bauman et al. [76] identified four prognostic groups of patients with statistically different median survivals (MS): (1) [KPS <70 and age >40y, MS: 12 m], (2) [KPS ≥70, age >40y, and enhancement present, MS: 46 m], (3) [KPS <70 and age: 18–40y, or KPS ≥70 and age >40y, no enhancement, MS: 87 m], and (4) [KPS ≥70 and age: 18–40y, MS: 128 m], with the following abbreviations: KPS Karnofsky performance Status, y years, m months.

In 22844 and 22845 EORTC trials, Pignatti et al. [64] showed that age >40 years, astrocytoma histology subtype, largest diameter of the tumor >6 cm, tumor crossing the midline, and presence of neurologic deficit before surgery were unfavorable prognostic factors for survival. The total number of unfavorable factors can be used to determine the prognostic score.

In the University of California at San Francisco LGG prognostic scoring system, patients were assigned a prognostic score based upon the sum of points assigned to the presence of each of the four following factors: (1) location of tumor in presumed eloquent cortex, (2) KPS Score ≤80, (3) age >50 years, and (4) maximum diameter >4 cm [77, 78]. Survival estimates according to this DLGG score were applied in 537 patients and are exposed in Tables 2.8 and 2.9.