3. All possible 2 × 2 test comparisons from the included studies.
| Test categories | aCGH | CGH | CISH | FISH | G‐banding | Methylation array | MLPA | MS | NanoString | NGS | PCR‐based LOH | RFLP | Real‐time PCR | SNP array |
| aCGH | — | — | — | Blesa 2009; Byeon 2014; Mohapatra 2006; Pesenti 2017 | — | — | Blesa 2009 | Pesenti 2017 | — | — | Blesa 2009; Cowell 2004; Mohapatra 2006 | — | — | — |
| CGH | — | — | — | Burger 2001; Hatanpaa 2003a (assay development and non‐blinded validation cohort); Hatanpaa 2003b (blinded validation cohort); Smith 1999 | Dahlback 2009; Dahlback 2011; Schrock 1994 | — | Jeuken 2006 | — | — | — | Bigner 1999; Burger 2001; Dahlback 2011; Hatanpaa 2003a (assay development and non‐blinded validation cohort); Hatanpaa 2003b (blinded validation cohort); Smith 1999 | — | — | — |
| CISH | — | — | — | Lass 2013 | — | — | — | — | — | — | — | — | — | — |
| FISH | — | — | — | Belaud‐Rotureau 2006a; Duval 2014b; Duval 2015c; Horbinski 2012d; Senetta 2013e; Srebotnik‐Kirbis 2016f; Uchida 2019g | — | — | Blesa 2009; Natte 2005 | Pesenti 2017 | Armanious 2017 | D'Haene 2019; Kato 2019; Na 2019; Park 2019; Sim 2018a (glioblastoma cohort)h; Sim 2018b (oligodendroglial cohort); Thomas 2017 | Blesa 2009; Bouvier 2004; Broholm 2008; Burger 2001; Cieply 2004; Clark 2013; Gadji 2009; Hatanpaa 2003a (assay development and non‐blinded validation cohort); Hatanpaa 2003b (blinded validation cohort); Horbinski 2012; Jha 2011; Mohapatra 2006; Pesenti 2017; Scheie 2006; Smith 1999 | — | Chaturbedi 2012; Nigro 2001 | Ghasimi 2016; Hinrichs 2016; Kolhe 2016; Lhotska 2015; Paxton 2015; Thakur 2012 |
| G‐banding | — | — | — | — | — | — | — | — | — | — | Dahlback 2011 | Ransom 1992a; Ransom 1992b | — | — |
| Methylation array | — | — | — | — | — | — | Wiestler 2014 | — | — | — | — | — | — | — |
| MLPA | — | — | — | — | — | — | — | — | — | — | Blesa 2009 | — | — | — |
| MS | — | — | — | — | — | — | — | — | — | — | Pesenti 2017 | — | — | — |
| NanoString | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| NGS | — | — | — | — | — | — | — | — | — | — | Dubbink 2016 | — | — | — |
| PCR‐based LOH | — | — | — | — | — | — | — | — | — | — | Hatanpaa 2003a (assay development and non‐blinded validation cohorti); Hatanpaa 2003b (blinded validation cohorti) | — | Ariza 2010 | Harada 2011; Tsiatis 2010 |
| RFLP | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| Real time PCR | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| SNP array | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
aCGH: array comparative genomic hybridisation; CGH: comparative genomic hybridisation; CISH: chromogenic in situ hybridisation; FISH: fluorescence in situ hybridisation; LOH: loss of heterozygosity; MLPA: multiplex‐ligation‐dependent probe amplification; NGS: next‐generation sequencing; PCR: polymerase chain reaction; RFLP: restriction fragment length polymorphism; SNP: single nucleotide polymorphism.
aBelaud‐Rotureau 2006 performed FISH with manual analysis with the 1p36.3 (D1Z2)/1q12 (D1Z1) and 19q13.3/19pter probe set, manual analysis with the 1p36/1q25 and 19q13/19p13 Abbott Vysis probe set, and automatic analysis (Metafer 4, Metasystems, Althlussheim, Germany) with the 1p36/1q25 and 19q13/19p13 Abbott Vysis probe set. bDuval 2014 performed FISH and immunoFISH (FISH with immunohistochemistry against Ki67 (MIB‐1)), and used two different cut‐offs for both – a "combination" cut‐off (which was based on the number of cells showing a deletion) and a "ratio" cut‐off (based on the ratio of signals for 1p to 1q and 19q and 19p). cDuval 2015 performed FISH with automated analysis (Metafer 4 software (Metasystem) using the "1p19q tile‐sampling classifier") and FISH with manual analysis. dHorbinski 2012 performed FISH with two different cut‐offs in addition to PCR‐based LOH (target‐ploidy control ratio < 0.87, with ≥ 20% of nuclei showing deletion and target‐ploidy control ratio < 0.75, with ≥ 20% of nuclei showing deletion). eSenetta 2013 performed FISH with two different cut‐offs (cut‐off ratios 1p ≤ 0.8 and 19q ≤ 0.8 and cut‐off ratios 1p ≤ 0.7 and 19q ≤ 0.8). fSrebotnik‐Kirbis 2016 FISH variants on fresh tissue cytospins and on FFPE sections. gUchida 2019 performed FISH with two different criteria for judging whether a deletion was present (signals of 1p or 19q < signals of 1q or 19p or single signal of 1q or 19q and two signals of 1q or 19p; in both cases the cut‐off value was set at 20%). hIn Sim 2018a and Sim 2018b, glioblastoma cohort FISH was compared to NGS or aCGH (or both). We categorised NGS or aCGH (or both) as NGS for the purposes of this table. iHatanpaa 2003a and Hatanpaa 2003b performed PCR‐based LOH with or without comparison to normal DNA (in addition to CGH and FISH). This study developed a cut‐off for PCR‐based LOH without comparison to normal DNA in one set of participants (Hatanpaa 2003a assay development and non‐blinded validation cohort) and validated it in another set of participants (Hatanpaa 2003b blinded validation cohort).