Keywords
hemoglobin–albumin–lymphocyte–platelet score - germinoma - germ cell tumor - pineal
tumor - neutrophil-to-lymphocyte ratio - prognosis - prognostic nutritional index
Introduction
Pineal tumors are rare, comprising approximately 0.4 to 1% of all malignant brain
tumors. These tumors account for 1 to 4% of brain tumors in adults and 10% of brain
tumors in children.[1]
[2] The pineal region contains a wide variety of histoanatomical structures, leading
to the appearance of various malignancies in this location, including germ cell tumors,
pineal parenchymal tumors, gliomas, miscellaneous tumors, and cysts.[3] Therefore, the management of malignancies in the pineal region is determined by
the specific histological subtypes.[4]
Prognostic factors for tumors of the pineal region have been identified through a
review of the relevant literature. According to Valsechi et al, the predictor of a
poor prognosis in germinoma was female gender, whereas male was associated with an
unfavorable prognosis in pineoblastoma.[5] Additionally, Cavalheiro et al found that entire resection was a potential predictor
of outcome in pediatric patients with pineal area tumors.[6] In a population-based study, older age, male gender, nongerm cell tumor, and receiving
chemotherapy were significantly associated with prognosis.[7]
In recent studies, several immunonutritional indexes that function as prognostic factors
for a wide range of malignancies have been identified.[8]
[9] Preoperative hemoglobin–albumin–lymphocyte–platelet (HALP) score of more than 24
was associated with disease recurrence and mortality in univariate analysis, while
the prognostic nutritional index (PNI) of more than 45 was found to be substantially
associated with mortality in endometrial cancer, according to Njoku et al.[10] Additionally, Leetanaporn and Hanprasertpong discovered that patients with locally
advanced cervical cancer who had a HALP score of more than 22.2 had a longer progressive-free
survival and overall survival time.[11] Shen et al discovered that HALP scores of more than 25.8 were substantially associated
with recurrence-free survival in elderly patients with gastric cancer or small-cell
lung cancer.[12] Furthermore, as a prognostic factor, the neutrophil-to-lymphocyte ratio (NLR) is
among the immune-inflammatory biomarkers that have been investigated. In univariate
analysis, Garrett et al demonstrated that NLR less than 5.07 and PNI more than 46.97
were significantly associated with prolonged survival time in univariate analysis;
however, in multivariate analysis, these biomarkers were not significantly associated
with survival.[13] Various immunonutritional indexes associated with long survival time that are potentially
explained by the indexes can represent patient status, both nutritional and inflammatory
status, and reflect the severity of the disease.[14]
[15]
According to a study of the literature, a few studies have been conducted on the prognostic
value of immunonutritional indexes in patients with intracranial tumors, particularly
pineal region tumors. In response to this knowledge gap, the authors aimed to assess
the prognostic value of HALP score, PNI, and NLR in patients with pineal region tumors.
Methods
Study Design and Study Population
After institutional ethics board approval (REC.65-431-10-1), a retrospective cohort
study of patients with pineal tumors who had confirmed diagnoses from tissue specimens
by a pathologist between January 2010 and December 2022 was conducted. Patients who
lacked complete medical records or histological slides for diagnosis confirmation,
or who could not identify their current state, were eliminated. Baseline patient information,
clinicopathological data, and preoperative laboratories were gathered such as age,
gender, extent of resection, tumor size, preoperative hydrocephalus, pathological
result, complete blood count, and tumor markers.
Before the review of the medical records, the operational definition was made. Based
on the research conducted by previous studies,[16]
[17] the evaluation of residual tumor and the extent of resection were performed utilizing
postoperative T1-weighted with contrast imaging. In detail, a postoperative residual
tumor of no more than 5% was referred to as gross total resection (GTR). Subtotal
resection was defined as a very small and usually strongly adhesive remnant portion
adhered to the venous walls or infiltrating important structures. Partial resection
was defined as the resection of less than 95% of the residual tumor that was visible
on postoperative imaging. Moreover, the biopsy was just used for diagnostic purposes
and not for tumor removal.[16]
[17]
[18]
The preoperative hematological measures comprised hemoglobin, leukocyte, neutrophil,
lymphocyte, platelet counts, and albumin that were used to determine HALP score, PNI,
and NLR. The HALP score was calculated using the following formula: hemoglobin (g/L) × albumin
(g/L) × lymphocyte (/L)/platelet (/L).[15] The PNI was determined as 10 × serum albumin (g/dL) + 0.005 × total lymphocyte count
(per mm3), while the ratio of neutrophil to lymphocyte counts was used to calculate NLR.[19]
For outcome evaluation, follow-up data were gathered until December 2023, which included
status information (death or staying alive). Follow-up data were gathered during patients'
visits to outpatient clinics or through information obtained from the death records
by the local municipality.
Statistical Analysis
Descriptive statistics were utilized to determine clinical characteristics and imaging
results. Mean and standard deviation were utilized for continuous variables, and percentages
were used for categorical data. For survival analysis, the median overall survival
time was estimated. Using maximum selected rank statistical techniques, the cutoff
threshold for each immunonutritional indicator was determined. The Cox proportional
hazard regression model was used to explore prognosis in both univariate and multivariable
analysis. Candidate variables with a p-value of 0.1 or less in the univariate analysis were examined in a multivariable
model using backward stepwise selection in order to finalize the final model. Therefore,
the model that obtained the lowest Akaike information criterion (AIC) value was selected
as the final model.
Thus, a Kaplan–Meier survival curve was established, and log-rank tests were utilized
to evaluate each predictor in the final model. Statistical analyses were performed
by the Jamovi version 2.3.28 software (The Jamovi project, Sydney, Australia).
Ethical Considerations
The study was conducted in accordance with the Declaration of Helsinki (as revised
in 2013). A human research ethics committee of the Faculty of Medicine, Prince of
Songkla University, Songkhla, Thailand approved the present study (REC.65-431-10-1).
Because it was a retrospective analysis, the current study did not require patients'
informed consent. However, the identity numbers of patients were encoded before analysis.
Results
Characteristics of Patients in the Present Cohort
The clinical characteristics of the 51 patients with pineal region tumors are shown
in [Table 1]. The majority of patients were males, with an average age of 20.3 ± 16.2 years.
Typical symptoms included worsening headache, alteration of consciousness, and ataxia
resulting from hydrocephalus before surgery. Additionally, preoperative hydrocephalus
was found in 84.3% of cases. Germinoma is the most common tumor in the pineal region,
accounting for 33.3% of cases, while pineal parenchymal tumors occurred in 7.8% of
cases. The study found an average preoperative tumor diameter of 3.64 ± 1.35 cm and
a mean tumor volume of 31.3 ± 33.5 mL.
Table 1
Baseline clinical characteristics
|
Factor
|
N (%)
|
|
Gender
|
|
|
Male
|
40 (78.4)
|
|
Female
|
11 (21.6)
|
|
Mean age, y (SD)
|
20.3 (16.2)
|
|
Age group, y
|
|
|
≤ 24
|
36 (70.6)
|
|
> 24
|
15 (29.4)
|
|
Preoperative symptoms
|
|
|
Headache
|
36 (70.6)
|
|
Alteration of consciousness
|
18 (35.3)
|
|
Ataxia
|
15 (29.4)
|
|
Weakness
|
11 (21.6)
|
|
Seizure
|
8 (15.7)
|
|
Behavior change
|
2 (3.9)
|
|
Glasgow Coma Score
|
|
|
< 13
|
5 (9.8)
|
|
≥ 13
|
46 (90.2)
|
|
Extent of resection
|
|
|
Gross total resection
|
5 (9.8)
|
|
Subtotal resection
|
7 (13.7)
|
|
Partial resection
|
15 (29.4)
|
|
Biopsy
|
24 (47.1)
|
|
Pathological result
|
|
|
Germinoma
|
17 (33.3)
|
|
Germ cell tumor (nongerminoma)
|
8 (15.7)
|
|
Nongerm cell tumor
|
26 (51.0)
|
|
Pilocytic astrocytoma
|
1 (2.0)
|
|
Diffuse astrocytoma
|
2 (3.9)
|
|
Anaplastic astrocytoma
|
4 (7.8)
|
|
Ependymoma
|
1 (2.0)
|
|
Subependymal giant cell astrocytoma
|
1 (2.0)
|
|
Pineal parenchymal tumor
|
4 (7.8)
|
|
Meningioma
|
3 (5.9)
|
|
Craniopharyngioma
|
1 (2.0)
|
|
Metastasis
|
1 (2.0)
|
|
Neuroblastoma
|
1 (2.0)
|
|
Other
|
7 (13.6)
|
Abbreviation: SD, standard deviation.
Preoperative laboratories, biomarkers, and immunonutritional indexes are shown in
[Table 2]. The mean serum alpha-fetoprotein was 215.00 ± 594.66 ng/mL, while the mean alpha-fetoprotein
in cerebrospinal fluid (CSF) was 183.39 ± 632.72 ng/mL. Additionally, human chorionic
gonadotropin (hCG) levels in the serum were mean of 74.12 ± 233.65 mIU/mL, while hCG
levels in the CSF were mean of 36.59 ± 87.04 mIU/mL. Additionally, the average HALP
score was 42.1 ± 29.6, and the mean PNI was 68.9 ± 16.4.
Table 2
Imaging findings and preoperative laboratory results
|
Factor
|
N (%)
|
|
Imaging finding
|
|
|
Preoperative hydrocephalus
|
43 (84.3)
|
|
Number of tumor
|
|
|
Single
|
44 (86.3)
|
|
Multiple
|
7 (13.7)
|
|
Mean midline shift, mm (SD)
|
0.412 (1.88)
|
|
Mean tumor diameter, cm (SD)
|
3.64 (1.35)
|
|
Mean tumor volume, mL3 (SD)
|
31.3 (33.5)
|
|
Preoperative laboratory
|
|
|
Mean hemoglobin, mg% (SD)
|
12.9 (1.98)
|
|
Mean lymphocyte count, cell/uL (SD)
|
2,586 (1612)
|
|
Mean platelet count, cell/uL (SD)
|
358,137 (115,309)
|
|
Mean albumin, g/dL (SD)
|
4.31 (0.457)
|
|
Biomarkers
|
|
|
Mean serum alpha-fetoprotein, ng/mL
|
215.00 (594.66)
|
|
Mean CSF alpha-fetoprotein, ng/mL
|
183.39 (632.72)
|
|
Mean serum beta hCG, mIU/mL
|
74.12 (233.65)
|
|
Mean CSF beta hCG, mIU/mL
|
36.59 (87.04)
|
|
Immune-nutritional indexes
|
|
|
Mean hemoglobin, albumin, lymphocyte, and platelet score (SD)
|
42.1 (29.6)
|
|
Mean neutrophil-to-lymphocyte ratio (SD)
|
49.9 (54.7)
|
|
Mean prognostic nutritional index (SD)
|
68.9 (16.4)
|
Abbreviations: CSF, cerebrospinal fluid; hCG, human chorionic gonadotropin; SD, standard
deviation.
Survival Analysis
Average follow-up time was 64.84 ± 45.00 months and 1-, 2-, and 5-year survival probabilities
were 92% (95% confidence interval [CI] 85–99.8), 92% (95% CI 85–99.8), and 81% (95%
CI 70–94.1), respectively. Moreover, the median overall survival time was determined
but the estimated median survival was not reached, as shown in [Fig. 1]. The extent of resection was estimated for association with prognosis, and we found
that the GTR group was not significantly associated with a favorable prognosis compared
with the non-GTR group (p = 0.33), as shown in [Fig. 2A]. Furthermore, [Fig. 2B] shows that the prognosis of patients with germinoma was not substantially different
from that of the nongerminoma group (p = 0.64). Additionally, 1-, 2-, and 5-year survival probabilities of germinoma were
88.2% (95% CI 74.2–100), 88.2% (95% CI 74.2–100), and 80.8% (95% CI 63.4–100), respectively.
Fig. 1 Kaplan–Meier survival curve of overall survival.
Fig. 2 Kaplan–Meier survival curves of the extent of resection and histological types of
tumor. (A) Gross total resection (GTR) group and non-GTR group. (B) Germinoma group and nongerminoma group.
In immunonutritional indexes, the maximum selected rank statistical method was used
to examine the HALP score, PNI, and NLP cutoff thresholds. These values were 22.6,
56.08, and 45.9, respectively. The Cox proportional hazard regression was performed
to explore factors associated with prognosis. Seizure symptom, HALP score, and PNI
were candidate variables in univariate analysis, as shown in [Table 3]. Kaplan–Meier survival curves of immunonutritional indexes and candidate variables
are demonstrated in [Fig. 3A–D]. Thus, these candidates were analyzed in multivariable analysis using a backward
stepwise procedure with the lowest AIC value. The HALP score was the only predictor
included in the final model (hazard ratio 0.25, 95% CI 0.06–1.00, p = 0.05). In addition, there was a significant difference in prognosis between the
high and low HALP score group by log-rank test (p = 0.035).
Fig. 3 Kaplan–Meier survival curves of immune-nutritional indexes and candidate variables.
(A) Hemoglobin–albumin–lymphocyte–platelet (HALP) score. (B) Prognostic nutritional index (PNI). (C) Neutrophil-to-lymphocyte ratio (NLR). (D) Preoperative seizure.
Discussion
The present study revealed a favorable prognosis for the pineal tumor region. The
5-year survival probability was 81%, which was marginally higher than the prognosis
from previous research. From 1973 to 2005, Al-Hussaini et al analyzed the prognosis
of 633 patients with pineal tumors using the Surveillance, Epidemiology, and End Results
(SEER) database. They discovered that the overall 5-year survival rate was 65%.[20] Additionally, Vuong et al used the SEER database from 1975 to 2016 to analyze the
survival of pineal tumors and the separated cases diagnosed between 1975 and 2016
into four groups for survival trend analysis: 1975 to 1984, 1985 to 1994, 1995 to
2004, and 2005 to 2016. Consequently, there has been a rise in the survival rate of
pineal gland tumors over time, which may be attributed to notable advancements in
technology and health care.[7] For pineal germinoma, the 5-year survival of germinoma in the present study was
80.8% which is consistent with prior studies. Cavalheiro et al reported that the overall
survival rate for patients with germinomas at 1, 2, and 5 years was 93.7, 93.7, and
88%, respectively, whereas Valsechi et al determined that the 5-year survival rate
for germinomas was 88%.[5]
[6]
Previous studies have identified prognostic factors for pineal tumors, including gender,
age, and the extent of resection. In the present study, the HALP score is strongly
associated with prognosis. In detail, a high HALP score demonstrated a significant
survival benefit compared to low HALP scores. This might be explained by the fact
that the score measures both the patient's immune system and nutritional status.[21] Because the HALP score had never been reported in the pineal tumor region in previous
research, and there was no standard cutoff value for the HAP score based on a literature
review, we utilized a statistical method to identify an optimal cutoff value that
could be validated in the future.[22] From literature review, a limited number of previous studies have revealed that
HALP serves as a prognostic factor.
Tunthanathip and Oearsakul studied various immune-nutritional indexes for prognostication
in glioblastoma and found that HALP score of more than 32 was significantly associated
with prolong survival time.[23] Wang et al reported that the HALP cutoff threshold for endometrial cancer was 33.8
and that the group with lower HALP had worse survival outcomes compared to the group
with higher HALP.[24] In gastrointestinal stromal tumors, HALP was a prognostic factor with a threshold
value of 31.5.[25] Furthermore, the result of the present study tended to utilize PNI as a predictor
due to the biomarker's statistical significance in univariate analysis. Nevertheless,
there was no significant correlation between PNI and prognosis in the multivariable
analysis. Immunonutritional indicators, which require regular laboratory tests that
are often performed on patients, can serve as straightforward and adjustable predictors
to help clinicians enhance the prognosis of patients with immunonutritional deficiencies.[8]
To the best of the authors' knowledge, this was the first study to show a significant
relationship between HALP and the prognosis of patients diagnosed with pineal region
tumors. However, it was noted that the present study had certain limitations. First,
the authors considered a limited number of patients for survival analysis. This limitation
could be resolved through multicenter research, which could boost the number of study
population and solve the overfitting problem of the predictive model by validation
with unseen data.[26]
[27] Therefore, the study's retrospective design introduced biases in both selection
and information. However, the authors attempted to address the limitation through
the creation of an operational definition, inclusion, and exclusion criteria before
the assessment.[28]
[29] Furthermore, the present study just assessed immunonutritional indexes in the preoperative
phase. In the future, a longitudinal study with a repeated-measures design should
be conducted to investigate the association between immunonutritional indexes and
prognosis.[30]
Conclusion
The preoperative HALP score is an independent prognostic factor for patients diagnosed
with pineal region tumors. The prognostication of the immunonutritional indexes requires
to be validated in the future.
Table 3
Univariate analysis by Cox regression analysis
|
Factor
|
Hazard ratio (95% CI)
|
p-Value
|
|
Age, y
|
0.96 (0.89–1.02)
|
0.18
|
|
Age group, y
|
|
|
|
≤ 24
|
Ref
|
|
|
> 24
|
0.57 (0.17–4.87)
|
0.99
|
|
Gender
|
|
|
|
Female
|
Ref
|
|
|
Male
|
31.09 (0.02–4853.10)
|
0.35
|
|
Weakness
|
|
|
|
No
|
Ref
|
|
|
Yes
|
0.46 (0.06–3.79)
|
0.47
|
|
Headache
|
|
|
|
No
|
Ref
|
|
|
Yes
|
3.30 (0.41–26.84)
|
0.26
|
|
Seizure
|
|
|
|
No
|
Ref
|
|
|
Yes
|
3.27 (0.78–13.72)
|
0.10
|
|
Behavior change
|
|
|
|
No
|
Ref
|
|
|
Yes
|
0.67 (0.28–3.73)
|
0.98
|
|
Ataxia
|
|
|
|
No
|
Ref
|
|
|
Yes
|
0.95 (0.19–4.74)
|
0.95
|
|
Alteration of consciousness
|
|
|
|
No
|
Ref
|
|
|
Yes
|
3.04 (0.73–12.72)
|
0.12
|
|
Blurred vision
|
|
|
|
No
|
Ref
|
|
|
Yes
|
3.08 (0.74–12.93)
|
0.12
|
|
Extraocular movement deficit
|
|
|
|
No
|
Ref
|
|
|
Yes
|
3.31 (0.67–16.47)
|
0.14
|
|
Glasgow Coma Score
|
|
|
|
≥ 13
|
Ref
|
|
|
< 13
|
2.65 (0.53–13.20)
|
0.23
|
|
Extent of resection
|
|
|
|
Gross total resection
|
Ref
|
|
|
Nontotal resection
|
0.04 (0.01–887.68)
|
0.53
|
|
Pathological result
|
|
|
|
Germinoma
|
Ref
|
|
|
Nongerminoma
|
0.71 (0.17–2.98)
|
0.64
|
|
Preoperative hydrocephalus
|
|
|
|
No
|
Ref
|
|
|
Yes
|
0.59 (0.12–2.93)
|
0.51
|
|
Number of tumors
|
|
|
|
Multiple
|
Ref
|
|
|
Single
|
0.96 (0.12–7.85)
|
0.97
|
|
Tumor volume, mL3
|
|
|
|
≤ 20.16
|
Ref
|
|
|
> 20.16
|
0.34 (0.07–1.69)
|
0.18
|
|
Serum alpha-fetoprotein, ng/mL
|
|
|
|
≤ 1.68
|
Ref
|
|
|
> 1.68
|
1.10 (0.97–1.17)
|
0.58
|
|
CSF alpha-fetoprotein, ng/mL
|
|
|
|
≤ 0.60
|
Ref
|
|
|
> 0.60
|
1.07 (0.84–1.27)
|
0.87
|
|
Serum beta hCG, mIU/mL
|
|
|
|
≤ 50.0
|
Ref
|
|
|
> 50.0
|
1.17 (0.89–1.24)
|
0.95
|
|
CSF beta hCG, mIU/mL
|
|
|
|
≤ 5
|
Ref
|
|
|
> 5
|
1.01 (0.99–1.11)
|
0.59
|
|
HALP score
|
|
|
|
≤ 22.6
|
Ref
|
|
|
> 22.6
|
0.25 (0.06–1.00)
|
0.05
|
|
PNI
|
|
|
|
≤ 56.08
|
Ref
|
|
|
> 56.08
|
0.31 (0.08–1.25)
|
0.10
|
|
NLR
|
|
|
|
≤ 45.9
|
Ref
|
|
|
> 45.9
|
0.75 (0.15–3.75)
|
0.72
|
Abbreviation: CI, confidence interval; CSF, cerebrospinal fluid; HALP, hemoglobin–albumin–lymphocyte–platelet
score; hCG, human chorionic gonadotropin; NLR, neutrophil lymphocyte ratio; PNI, prognostic
nutritional index.
