On 2017 Aug 29, Marcus de Jong commented:

With great interest, but also with significant apprehension, we read the article by Yamanaka et al. recently published on-line ahead of print in Neurosurgical Review, entitled "Trilateral retinoblastoma: A systematic review of 211 cases."

We have had a long-time interest in retinoblastoma, including trilateral retinoblastoma, and we became worried about the fact that the authors allegedly have included 211 patients with trilateral retinoblastoma in their statistical analysis. We have reason to believe that this number, and consequently, the presented results may not be correct. In 2014, we published a systematic review and meta-analysis on trilateral retinoblastoma in Lancet Oncology, based on strict adherence to the PRISMA Statement (de Jong et al. [1]). We also contacted authors to resolve any equivocal issues. In particular, we meticulously matched patients between reports to prevent including any patients twice in our analysis. Many patients with trilateral retinoblastoma appear in the literature two or more times over the years; in the extreme case, one patient features in five different reports over 12 years. Resolving duplications and consolidating the sequential reports by each patient, we ended up with 174 unique individuals with trilateral retinoblastoma (see the attached table from the online supplement of our article).

Naturally, a number of trilateral retinoblastoma cases included in Yamanaka et al.'s paper have been published after acceptance of our paper. These are Andrade et al. [2] with 1 case, De Ioris et al. [3] with 2 new cases, and Pham et al. [4] with 3 cases, amounting to a total difference of 6 unique patients.

However, we included several articles that Yamanaka et al. have not ascertained during their research: De Jong et al. [5], Duncan et al. [6], Dunst et al. [7], Gururangan et al. [8], Jin et al. [9], Lim et al. [10], Mauger et al. [11], Onadim et al. [12], Popovic et al. [13], White et al. [14], and Zimmerman et al. [15]. These amount to a total of 18 unique cases.

We became very concerned after noting that, correcting for new cases and articles not included by Yamanaka et al., the difference in the number of patients between the two articles is no less than 49 (our 174 - 18 not included by Yamanaka et al. = 156 in De Jong et al. [1] versus their 211 - 6 newly reported = 205 in Yamanaka et al. [16]). Because Yamanaka et al. have not provided a patient-by-patient table of their cases unlike we did, we are limited to this numerical comparison and cannot verify case by case their unique cases.

In summary, we have reason to believe that Yamanaka et al. have not accounted for the fact that trilateral retinoblastoma patients have often been published in more than one paper. Although the results of the meta-analysis by Yamanaka et al., in general, show results that appear largely similar to those that what we published, the unexplained and large difference in the number of patients suggests to us that most, if not all, percentages and p-values from the meta-analysis by Yamanaka et al. do not reflect reality. There is also no way of knowing whether a particular statistic is correct or not.

To solve this issue, we ask Yamanaka et al. to clarify this crucial aspect of their meta-analysis, ideally by providing a patient-by-patient list of the 211 cases to verify their uniqueness to be attached to their paper. Hopefully the authors can provide a valid explanation that completely resolves our concerns. Otherwise, we are afraid that the statistical results of this paper will be misleading and cannot be trusted, and then perhaps the paper in the present form should be retracted in order to correct the literature and to avoid wrong interpretations in clinical practice of managing children with trilateral retinoblastoma.

Marcus de Jong,

on behalf of the authors of de Jong et al. [1]

References

1.de Jong MC, Kors WA, de Graaf P, et al (2014) Trilateral retinoblastoma: A systematic review and meta-analysis. Lancet Oncol 15:1157–1167. doi: 10.1016/S1470-2045(14)70336-5

2.Andrade GC de, Pinto NP de C, Motono M, et al (2015) Trilateral retinoblastoma with unilateral eye involvement. Rev Assoc Med Bras 61:308–10. doi: 10.1590/1806-9282.61.04.308

3.De Ioris MA, Valente P, Randisi F, et al (2014) Baseline central nervous system magnetic resonance imaging in early detection of trilateral retinoblastoma: pitfalls in the diagnosis of pineal gland lesions. Anticancer Res 34:7449–54.

4.Pham TTH, Siebert E, Asbach P, et al (2015) Magnetic resonance imaging based morphologic evaluation of the pineal gland for suspected pineoblastoma in retinoblastoma patients and age-matched controls. J Neurol Sci 359:185–192. doi: 10.1016/j.jns.2015.10.046

5.de Jong MC, Moll AC, Göricke S, et al (2016) From a Suspicious Cystic Pineal Gland to Pineoblastoma in a Patient with Familial Unilateral Retinoblastoma. Ophthalmic Genet 37:116–8. doi: 10.3109/13816810.2014.929717

6.Duncan JL, Scott IU, Murray TG, et al (2001) Routine neuroimaging in retinoblastoma for the detection of intracranial tumors. Arch Ophthalmol 119:450–2. 7.Dunst J, Fellner E, Erhardt J (1990) Trilaterales Retinoblastom mit spinalen Metastasen. Rofo 153:343–4. doi: 10.1055/s-2008-1033391

8.Gururangan S, McLaughlin C, Quinn J, et al (2003) High-dose chemotherapy with autologous stem-cell rescue in children and adults with newly diagnosed pineoblastomas. J Clin Oncol 21:2187–91. doi: 10.1200/JCO.2003.10.096

9.Jin J, Tang H-F, Zhou Y-B (2006) Trilateral retinoblastoma: a case report. World J Pediatr 2:151–153.

10.Lim FPM, Soh SY, Iyer JV, et al (2013) Clinical profile, management, and outcome of retinoblastoma in Singapore. J Pediatr Ophthalmol Strabismus 50:106–12.

11.Mauger TF, Makley TA, Davidorf FH, Rogers GL (1992) Retinoblastoma, microphthalmia, coloboma, and neuroepithelioma of the pineal body. Ann Ophthalmol 24:290–4.

12.Onadim Z, Woolford AJ, Kingston JE, Hungerford JL (1997) The RB1 gene mutation in a child with ectopic intracranial retinoblastoma. Br J Cancer 76:1405–9.

13.Popovic MB, Balmer A, Maeder P, et al (2006) Benign pineal cysts in children with bilateral retinoblastoma: a new variant of trilateral retinoblastoma? Pediatr Blood Cancer 46:755–61.

14.White L, Johnston H, Jones R, et al (1993) Postoperative chemotherapy without radiation in young children with malignant non-astrocytic brain tumours. A report from the Australia and New Zealand Childhood Cancer Study Group (ANZCCSG). Cancer Chemother Pharmacol 32:403–6.

15.Zimmerman L (1985) Trilateral retinoblastoma. In: Blodi F (ed) Retinoblastoma. Churchill Livingstone, New York, pp 185–210

16.Yamanaka R, Hayano A, Takashima Y (2017) Trilateral retinoblastoma: A systematic review of 211 cases. Neurosurg Rev. doi: 10.1007/s10143-017-0890-4

Also posted here: https://pubpeer.com/publications/50D5D3EFEA81DCD2421841D84FDA8D#

On 2017 Aug 29, Marcus de Jong commented:

With great interest, but also with significant apprehension, we read the article by Yamanaka et al. recently published on-line ahead of print in Neurosurgical Review, entitled "Trilateral retinoblastoma: A systematic review of 211 cases."

We have had a long-time interest in retinoblastoma, including trilateral retinoblastoma, and we became worried about the fact that the authors allegedly have included 211 patients with trilateral retinoblastoma in their statistical analysis. We have reason to believe that this number, and consequently, the presented results may not be correct. In 2014, we published a systematic review and meta-analysis on trilateral retinoblastoma in Lancet Oncology, based on strict adherence to the PRISMA Statement (de Jong et al. [1]). We also contacted authors to resolve any equivocal issues. In particular, we meticulously matched patients between reports to prevent including any patients twice in our analysis. Many patients with trilateral retinoblastoma appear in the literature two or more times over the years; in the extreme case, one patient features in five different reports over 12 years. Resolving duplications and consolidating the sequential reports by each patient, we ended up with 174 unique individuals with trilateral retinoblastoma (see the attached table from the online supplement of our article).

Naturally, a number of trilateral retinoblastoma cases included in Yamanaka et al.'s paper have been published after acceptance of our paper. These are Andrade et al. [2] with 1 case, De Ioris et al. [3] with 2 new cases, and Pham et al. [4] with 3 cases, amounting to a total difference of 6 unique patients.

However, we included several articles that Yamanaka et al. have not ascertained during their research: De Jong et al. [5], Duncan et al. [6], Dunst et al. [7], Gururangan et al. [8], Jin et al. [9], Lim et al. [10], Mauger et al. [11], Onadim et al. [12], Popovic et al. [13], White et al. [14], and Zimmerman et al. [15]. These amount to a total of 18 unique cases.

We became very concerned after noting that, correcting for new cases and articles not included by Yamanaka et al., the difference in the number of patients between the two articles is no less than 49 (our 174 - 18 not included by Yamanaka et al. = 156 in De Jong et al. [1] versus their 211 - 6 newly reported = 205 in Yamanaka et al. [16]). Because Yamanaka et al. have not provided a patient-by-patient table of their cases unlike we did, we are limited to this numerical comparison and cannot verify case by case their unique cases.

In summary, we have reason to believe that Yamanaka et al. have not accounted for the fact that trilateral retinoblastoma patients have often been published in more than one paper. Although the results of the meta-analysis by Yamanaka et al., in general, show results that appear largely similar to those that what we published, the unexplained and large difference in the number of patients suggests to us that most, if not all, percentages and p-values from the meta-analysis by Yamanaka et al. do not reflect reality. There is also no way of knowing whether a particular statistic is correct or not.

To solve this issue, we ask Yamanaka et al. to clarify this crucial aspect of their meta-analysis, ideally by providing a patient-by-patient list of the 211 cases to verify their uniqueness to be attached to their paper. Hopefully the authors can provide a valid explanation that completely resolves our concerns. Otherwise, we are afraid that the statistical results of this paper will be misleading and cannot be trusted, and then perhaps the paper in the present form should be retracted in order to correct the literature and to avoid wrong interpretations in clinical practice of managing children with trilateral retinoblastoma.

Marcus de Jong,

on behalf of the authors of de Jong et al. [1]

References

1.de Jong MC, Kors WA, de Graaf P, et al (2014) Trilateral retinoblastoma: A systematic review and meta-analysis. Lancet Oncol 15:1157–1167. doi: 10.1016/S1470-2045(14)70336-5

2.Andrade GC de, Pinto NP de C, Motono M, et al (2015) Trilateral retinoblastoma with unilateral eye involvement. Rev Assoc Med Bras 61:308–10. doi: 10.1590/1806-9282.61.04.308

3.De Ioris MA, Valente P, Randisi F, et al (2014) Baseline central nervous system magnetic resonance imaging in early detection of trilateral retinoblastoma: pitfalls in the diagnosis of pineal gland lesions. Anticancer Res 34:7449–54.

4.Pham TTH, Siebert E, Asbach P, et al (2015) Magnetic resonance imaging based morphologic evaluation of the pineal gland for suspected pineoblastoma in retinoblastoma patients and age-matched controls. J Neurol Sci 359:185–192. doi: 10.1016/j.jns.2015.10.046

5.de Jong MC, Moll AC, Göricke S, et al (2016) From a Suspicious Cystic Pineal Gland to Pineoblastoma in a Patient with Familial Unilateral Retinoblastoma. Ophthalmic Genet 37:116–8. doi: 10.3109/13816810.2014.929717

6.Duncan JL, Scott IU, Murray TG, et al (2001) Routine neuroimaging in retinoblastoma for the detection of intracranial tumors. Arch Ophthalmol 119:450–2. 7.Dunst J, Fellner E, Erhardt J (1990) Trilaterales Retinoblastom mit spinalen Metastasen. Rofo 153:343–4. doi: 10.1055/s-2008-1033391

8.Gururangan S, McLaughlin C, Quinn J, et al (2003) High-dose chemotherapy with autologous stem-cell rescue in children and adults with newly diagnosed pineoblastomas. J Clin Oncol 21:2187–91. doi: 10.1200/JCO.2003.10.096

9.Jin J, Tang H-F, Zhou Y-B (2006) Trilateral retinoblastoma: a case report. World J Pediatr 2:151–153.

10.Lim FPM, Soh SY, Iyer JV, et al (2013) Clinical profile, management, and outcome of retinoblastoma in Singapore. J Pediatr Ophthalmol Strabismus 50:106–12.

11.Mauger TF, Makley TA, Davidorf FH, Rogers GL (1992) Retinoblastoma, microphthalmia, coloboma, and neuroepithelioma of the pineal body. Ann Ophthalmol 24:290–4.

12.Onadim Z, Woolford AJ, Kingston JE, Hungerford JL (1997) The RB1 gene mutation in a child with ectopic intracranial retinoblastoma. Br J Cancer 76:1405–9.

13.Popovic MB, Balmer A, Maeder P, et al (2006) Benign pineal cysts in children with bilateral retinoblastoma: a new variant of trilateral retinoblastoma? Pediatr Blood Cancer 46:755–61.

14.White L, Johnston H, Jones R, et al (1993) Postoperative chemotherapy without radiation in young children with malignant non-astrocytic brain tumours. A report from the Australia and New Zealand Childhood Cancer Study Group (ANZCCSG). Cancer Chemother Pharmacol 32:403–6.

15.Zimmerman L (1985) Trilateral retinoblastoma. In: Blodi F (ed) Retinoblastoma. Churchill Livingstone, New York, pp 185–210

16.Yamanaka R, Hayano A, Takashima Y (2017) Trilateral retinoblastoma: A systematic review of 211 cases. Neurosurg Rev. doi: 10.1007/s10143-017-0890-4

Also posted here: https://pubpeer.com/publications/50D5D3EFEA81DCD2421841D84FDA8D#