The study used array comparative genomic hybridization and SNP-array records generated with two platforms over the study period. All software calls were manually reviewed, and potential findings were retained when they were at least 100 kilobases, involved at least 20 consecutive probes, or overlapped a known disease-associated gene or recurrent region. The researchers then re-curated all records against a single April 28, 2026 ClinGen release and converted earlier GRCh37 coordinates to GRCh38. This approach was designed to avoid counting partial overlaps, terminal abnormalities, complex rearrangements or unusually sized events as canonical recurrent CNVs merely because they touched a recognized locus.

To assign a recurrent region, investigators required agreement between the array-defined breakpoints and the flanking low-copy repeat blocks, together with a size compatible with the canonical interval. Minor endpoint differences were accepted when they were consistent with the same established region, but a percentage-overlap threshold was not used. Classification was a separate step: deletions were evaluated against region-specific haploinsufficiency evidence and duplications against triplosensitivity evidence under the 2019 ACMG/ClinGen standards. This distinction was central to the results, because recurrence describes how a copy-number change arises, whereas pathogenicity depends on its exact region, dosage direction and supporting evidence.

Overall, a pathogenic or likely pathogenic copy-number finding was identified in 278 of 1,661 postnatal probands, or 16.74%, and in 28 of 401 prenatal samples, or 6.98%. The study identified 109 recurrent-CNV carriers, including one postnatal proband with two independent recurrent events. Among the 92 postnatal carriers, 66 had pathogenic or likely pathogenic recurrent CNVs, representing 23.74% of all postnatal pathogenic copy-number findings. Seven of 17 prenatal carriers had pathogenic recurrent CNVs, accounting for 25.00% of prenatal pathogenic findings. The remaining recurrent events were classified as variants of uncertain significance or benign rather than being treated as diagnostically equivalent to pathogenic events.

The regional distribution varied between the two clinical settings. Proximal 22q11.21 deletions, associated with DiGeorge or velocardiofacial syndrome, and proximal 16p11.2 CNVs were the most frequent postnatal findings. Other recurrent entities included Xp22.31, proximal 7q11.23, 17q12 and 15q11.2–q13.1. Deletions predominated after birth, accounting for 64 of 93 postnatal recurrent-CNV events, or 68.82%. By contrast, duplications accounted for 11 of 17 prenatal events, or 64.71%. The prenatal pathogenic fraction was correspondingly lower, underscoring that the dosage direction of a recurrent CNV cannot be inferred from the region label alone.

Test purpose strongly influenced the observed yield among postnatal probands. Pathogenic recurrent CNVs were found in six of 19 cases tested because a specific recurrent-CNV syndrome was suspected, a yield of 31.58%, although this small subgroup had a wide 95% confidence interval of 15.36% to 53.99%. Among 1,589 phenotype-driven diagnostic tests without prior suspicion of a particular recurrent syndrome, 58 carried a pathogenic recurrent CNV, giving a yield of 3.65%. The targeted group produced the higher proportion, but phenotype-driven testing contributed 58 of the 66 postnatal pathogenic recurrent-CNV carriers. Neurodevelopmental and neurological referrals were the largest postnatal referral category and accounted for 38 of those 66 carriers.

Most fetal samples were referred because of an ultrasound abnormality, which was the primary indication in 348 of 401 cases, or 86.78%. Within this subgroup, 12 fetuses carried a recurrent CNV and six carried a pathogenic recurrent CNV, corresponding to 3.45% and 1.72%, respectively. The investigators described several concordant examples, including a 22q11.2 deletion with conotruncal and nuchal findings, a distal 1q21.1 deletion with left-sided cardiac hypoplasia and a 17q12/HNF1B deletion associated with a renal phenotype. They also reported discordant findings, such as a proximal 16p11.2 deletion in a fetus with myelomeningocele and Chiari II malformation, emphasizing that a pathogenic classification does not by itself establish that a CNV explains an ultrasound phenotype.

Inheritance added another layer to interpretation. Among 46 recurrent-CNV carriers for whom parental information was available, only one transmitting parent had a documented clinical diagnosis; the remaining information came from referral records rather than systematic parental examination. All 11 prenatal transmitting parents were recorded as clinically unaffected. In the prenatal cohort, five of six de novo findings were pathogenic, compared with two of 11 inherited findings, while nine inherited events were variants of uncertain significance or benign. These numbers were small and descriptive, but they illustrate the difficulty of predicting fetal outcomes from recurrence and inheritance alone, particularly for low-penetrance duplications and other regions with variable expressivity.

The authors said the results support region- and dosage-specific reporting of recurrent CNVs, with ClinGen reference coordinates, copy-number direction, ACMG/ClinGen classification, inheritance and phenotype concordance considered separately. This would help distinguish adjacent but clinically different intervals, such as proximal and distal 16p11.2 or the several recurrent regions around 22q11.21. The analysis was limited by its retrospective single-center design, unequal cohort sizes and periods, referral-based ascertainment, small subgroups and wide confidence intervals. It also excluded non-recurrent pathogenic CNVs, used two array platforms with different resolutions and reporting thresholds, and lacked systematic follow-up of pregnancy and postnatal outcomes. Consequently, its frequencies describe a clinical microarray referral population rather than population prevalence, and the authors did not estimate penetrance from the cohort.

Subject of Research: ClinGen-based classification of recurrent copy-number variants in prenatal and postnatal chromosomal microarray testing

Article Title: Recurrent CNVs in constitutional CMA: a ClinGen-anchored single-center postnatal and prenatal experience

Article References: Satkın, N. B., Atlı, E., Yalcintepe, S., Çelik, K. M., & Gürkan, H. (2026). Recurrent CNVs in constitutional CMA: a ClinGen-anchored single-center postnatal and prenatal experience. Frontiers in Genetics, 17, Article 1960791. https://doi.org/10.3389/fgene.2026.1960791

Image Credits: AI Generated

DOI: 10.3389/fgene.2026.1960791

Keywords: recurrent copy-number variants, chromosomal microarray, ClinGen, ACMG classification, prenatal diagnosis, postnatal diagnosis, dosage sensitivity, non-allelic homologous recombination, cytogenomics, Recurrent, CNVs, constitutional

The research team, led by Chuan Zhao and Junping Zhang of the Department of Neurosurgery and Neuro-oncology at Sanbo Brain Hospital, Capital Medical University, analyzed the records of 49 patients treated between 2016 and 2023 who had experienced at least three recurrences of an intracranial germ cell tumor. The cohort was overwhelmingly male, with 39 boys and men (79.6 percent) and 10 girls and women (20.4 percent), reflecting the known male predominance of this tumor type. The mean age at initial presentation was just 13.29 years, with a standard deviation of 5.99 years, underscoring that these are overwhelmingly diseases of children and adolescents. Histologically, 13 patients (26.5 percent) had pure germinomas, while 36 patients (73.5 percent) had non-germinomatous germ cell tumors, or NGGCTs, a heterogeneous group that includes teratomas, yolk sac tumors, choriocarcinomas, and embryonal carcinomas.

The number of recurrences ranged from 3 to 11, with most patients (61.2 percent) experiencing exactly three relapses. Ten patients (20.4 percent) relapsed four times, four patients (8.2 percent) five times, and five patients relapsed six or more times, including one child who endured an extraordinary 11 recurrences over more than a decade of treatment. The study recorded treatment modalities at initial diagnosis and at each recurrence, including surgical interventions, radiation therapy, chemotherapy regimens, and any associated adverse events. Recurrence was defined as the reappearance or progression of tumor after initial response, confirmed by new or enlarged lesions on contrast-enhanced MRI, histological confirmation of viable tumor, or a sustained rise in serum alpha-fetoprotein (AFP) or human chorionic gonadotropin (HCG) above the upper limit of normal.

One of the most striking findings of the study concerns the temporal pattern of relapse. The researchers defined progression-free survival (PFS) intervals between successive recurrences: PFS1 as the time from initial diagnosis to first recurrence, PFS2 as the interval from first to second recurrence, and PFS3 as the interval from second to third recurrence. For pure germinomas, the median PFS1 was 25.03 months, but the median PFS2 dropped to 14.16 months, a statistically significant shortening (P = 0.0286). The Friedman test, a non-parametric statistical test for detecting differences in treatments across multiple test attempts, revealed a highly significant decreasing trend in PFS intervals with successive recurrences in germinomas (P < 0.0001). In other words, once a germinoma begins to relapse, it tends to do so faster and faster.

The same pattern held for NGGCTs, but with even shorter intervals. The median PFS1 for NGGCT patients was 9.05 months, the median PFS2 was 5.61 months (P = 0.0428), and the median PFS3 was just 4.02 months (P < 0.0001). The Friedman test again demonstrated a statistically significant shortening of PFS intervals with successive relapses (P = 0.0092). This accelerating rhythm of recurrence is a critical clinical observation, because it suggests that multiply recurrent germ cell tumors become increasingly resistant to conventional salvage therapies over time, leaving clinicians with a progressively narrowing therapeutic window.

Overall survival told a similarly sobering story. The overall survival of germinoma patients was significantly better than that of NGGCT patients (P = 0.0019), consistent with the established biological distinction between these two tumor types. Germinomas are exquisitely sensitive to radiotherapy and chemotherapy and generally carry an excellent prognosis, with 5-year survival rates in the broader literature ranging from 70 to 90 percent. NGGCTs, by contrast, have variable and often less favorable outcomes due to their diverse histology and treatment responses. Yet the presence of 13 multiply recurrent germinoma patients in this cohort, accounting for 26.5 percent of the sample, is itself remarkable, because germinomas typically account for about two-thirds of intracranial germ cell tumors and rarely relapse after standard therapy. This suggests that a subset of germinomas may be intrinsically resistant to conventional radiotherapy and chemotherapy, a finding that the authors argue warrants exploration of alternative therapeutic strategies such as immunotherapy or targeted therapy guided by specific molecular characteristics.

Multivariate Cox proportional hazards regression analysis, performed to identify factors associated with progression-free survival before the first recurrence, revealed no statistically significant predictors in the germinoma subgroup for age at diagnosis, sex, tumor location, presence of dissemination, radiotherapy, chemotherapy, resection, or biomarkers. In the NGGCT subgroup, however, chemotherapy at initial diagnosis emerged as independently associated with shorter PFS1 (P < 0.001). The authors are careful to note that this result likely reflects confounding by indication rather than a detrimental effect of chemotherapy itself: in clinical practice, chemotherapy for NGGCT is rarely used alone but rather in conjunction with radiotherapy, and the majority of patients (69.4 percent, 25 of 36) underwent multimodal therapy. More intensive chemotherapy was likely selected for patients with higher-risk disease features at presentation, underscoring the importance of risk-adapted, multimodality therapy as established in protocols like SIOP-CNS-GCT-96.

The treatment trajectories documented in the study illustrate a shifting paradigm in the management of multiply recurrent disease. While initial management is multimodal, combining surgery, radiation, and chemotherapy, upon multiple recurrences the therapeutic arsenal narrows predominantly to systemic chemotherapy, as the feasibility of repeat surgery and radiotherapy diminishes due to cumulative toxicity and anatomical constraints. In the germinoma subgroup, all 13 patients received radiotherapy as part of primary treatment, but by the third recurrence only 4 of 11 patients treated received radiotherapy, while 11 of 11 received chemotherapy. Similarly, in the NGGCT subgroup, 22 of 36 patients underwent surgery at initial diagnosis, but by the second recurrence only 8 did, while 29 of 36 received chemotherapy. One patient received immunotherapy at the third recurrence of a germinoma, and one NGGCT patient received targeted therapy at the second recurrence, hinting at emerging salvage options beyond conventional cytotoxic agents.

The individual case histories in the study are sobering illustrations of the human cost of multiply recurrent disease. One 9-year-old girl presented in 2013 with a suprasellar mass and endured five recurrences involving intraspinal and intracranial dissemination over eight years, treated with repeated BEP (cisplatin, etoposide, bleomycin), spinal radiotherapy, ifosfamide-carboplatin-etoposide, pemetrexed-cisplatin with Endostar, whole-brain plus local radiotherapy, and temozolomide with anlotinib, before dying of disease progression in 2021 with an overall survival of 8 years and 1 month. A 16-year-old boy with a pineal germinoma experienced five recurrences with extensive intraspinal, ventricular, brainstem, and cerebellar dissemination, treated with multiple chemotherapy regimens and repeated radiotherapy; at last follow-up in 2023 he was alive without recurrence but had impaired balance and memory. A third patient, a boy with an NGGCT containing mature teratoma components, endured 11 recurrences over more than a decade, treated with surgery, whole-spinal and local radiotherapy, gamma knife radiosurgery, multiple chemotherapy protocols, and apatinib, before eventually dying of disease progression.

The authors acknowledge several limitations of their work. It is a retrospective study rather than a randomized trial, it lacks molecular-level and mechanistic investigations, and the cases were treated at multiple hospitals with various confounding factors, although this also reflects the real-world situation of repeatedly recurrent germ cell tumors. The absence of a non-recurrent control group and the relatively small sample size limit the ability to detect factors uniquely predictive of multiple relapses; based on the observed effect sizes, approximately 120 to 150 patients would be required for 80 percent statistical power. The lack of centralized pathology review and molecular profiling data prevents exploration of the biological determinants of aggressive, recurrent disease. The authors call for future multicenter prospective studies with matched controls, centralized pathology review, and molecular profiling, including methylation subtyping and driver gene analysis, to identify biological subtypes underlying the multiply recurrent phenotype. For now, the study provides a sobering quantitative foundation for understanding what happens when these rare brain tumors refuse to stay away: the intervals between relapses shorten with each successive recurrence, management increasingly relies on systemic chemotherapy as local modalities become less feasible, and NGGCT patients face significantly worse overall survival than those with germinomas.

Subject of Research: A retrospective study: multiply recurrent intracranial germ cell tumors

Article Title: A retrospective study: multiply recurrent intracranial germ cell tumors

Article References: Zhao, C., Ge, J., Li, C., Zhao, C., Qi, S., Xue, F., Zhang, J., & Zhang, J. (2026). A retrospective study: multiply recurrent intracranial germ cell tumors. Frontiers in Oncology, 16, Article 1962932. https://doi.org/10.3389/fonc.2026.1962932

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1962932

Keywords: neuro-oncology, germ cell tumors, brain tumor recurrence, pediatric cancer, chemotherapy, cancer research, retrospective, multiply, recurrent, intracranial, germ, cell

The strategy, which the authors call modified induction–maintenance therapy, or MIMT, departs from the conventional playbook in a fundamental way. Standard first-line regimens for recurrent or metastatic cervical cancer now combine chemotherapy with an immune checkpoint inhibitor and, in many protocols, the anti-angiogenic antibody bevacizumab, continued until disease progression. In the landmark KEYNOTE-826 trial, adding pembrolizumab to chemotherapy plus bevacizumab extended median overall survival to 28.6 months in patients whose tumors expressed PD-L1, compared with 16.5 months in controls. The phase 3 GOG 240 trial had earlier shown that bevacizumab alone pushed overall survival from 13.3 to 16.8 months. Yet these regimens carry a hidden cost: months of continuous triple-drug exposure accumulate toxicity, financial burden, and diminishing returns, particularly in patients whose tumors lack PD-L1 expression and who derive little survival advantage from checkpoint blockade.

MIMT flips the timing. Instead of continuing all three drug classes until progression, the Yunnan team delivered a deliberately short induction phase—just two to four cycles of platinum-based chemotherapy, bevacizumab, and a PD-1 inhibitor such as toripalimab or tislelizumab—then dropped the cytotoxic chemotherapy entirely once imaging showed disease control. What remained was a gentler dual maintenance of bevacizumab plus the checkpoint inhibitor, sustained for months or years. The logic is elegant: the induction phase acts as a biological trigger, rapidly debulking the tumor and releasing tumor antigens while the immune system is primed by checkpoint blockade, and the maintenance phase sustains that immune activation while starving the tumor of its vascular supply, all without the cumulative damage of prolonged chemotherapy.

The four cases tell a chronological story of escalating success. Case 1, a 53-year-old woman with FIGO stage IIB adenocarcinoma diagnosed in January 2020, developed peritoneal metastasis with large-volume ascites within a year. She received four cycles of albumin-bound paclitaxel, carboplatin, and bevacizumab, followed by five cycles of bevacizumab plus toripalimab. After two maintenance cycles, imaging showed a partial response, and cytological examination of her ascitic fluid revealed no malignant cells. She died in September 2023, approximately 44 months from diagnosis. Case 2, also 53, presented with stage IIIC1 squamous cell carcinoma and later developed vesicovaginal and rectovaginal fistulas along with bilateral hydronephrosis. Two cycles of triplet induction with tislelizumab produced stable disease, followed by ten cycles of dual maintenance that yielded a partial response. Her first progression-free interval lasted about 17 months; a second salvage regimen added another 7.5 months before she died in September 2023, roughly 43 months from diagnosis.

Case 3 is where the pattern becomes remarkable. A 57-year-old woman with stage IB2 squamous cell carcinoma diagnosed in August 2019 relapsed in July 2021 with retroperitoneal and iliac lymph node metastases causing left hydronephrosis. She received just two cycles of albumin-bound paclitaxel, nedaplatin, bevacizumab, and toripalimab, achieving a partial response, then transitioned to six cycles of bevacizumab plus the PD-1 inhibitor. When toripalimab became financially unsustainable, she continued bevacizumab alone with intermittent checkpoint inhibitor dosing. Her first progression-free interval stretched to 46 months before an axillary lymph node biopsy confirmed metastatic disease in June 2025. Reinduction with the four-drug combination produced another partial response, and her second progression-free interval has exceeded 10 months. Her overall survival from diagnosis has surpassed 83 months, and she remains alive under active follow-up as of July 2026.

Case 4 may be the most striking of all. A 59-year-old woman with recurrent cervical cancer complicated by a rectovaginal fistula arrived at the hospital in August 2021 with an Eastern Cooperative Oncology Group performance status of 3 and full cancer cachexia—so frail that triplet induction seemed inappropriate. The team proceeded anyway, delivering two cycles at 80 percent of standard chemotherapy doses. Grade 4 myelosuppression forced discontinuation after the second cycle, but after supportive care with granulocyte colony-stimulating factor and platelet transfusions, her blood counts recovered and her cachexia gradually improved. Switching to bevacizumab plus toripalimab maintenance, she has now been in remission for more than five years with no documented progression. Her progression-free survival exceeds 58 months and her overall survival stands at 86 months. From 2024 onward, her maintenance interval was stretched to once every two to three months, and she continues the dual regimen at a local hospital.

Two patients provided perspectives that capture the human weight of these numbers. Case 3 said she has lived with the disease for over seven years and is still here with her family, describing the treatment as giving her time she never expected to have. Case 4 recalled being unable to get out of bed or eat before treatment, and now being able to do the things she loves. These are not laboratory abstractions but women whose daily functioning was restored by a regimen that, on paper, involved far less chemotherapy than conventional protocols would prescribe.

The authors advance a hypothesis they describe as consistent with immune-mediated trigger-and-sustain effects. In this framework, the short triplet induction rapidly reduces tumor burden, releasing antigens and inflammatory signals that activate anti-tumor immunity, while the checkpoint inhibitor removes the brakes on T-cell function and bevacizumab normalizes the tumor vasculature, allowing immune cells to infiltrate more effectively. Once that immune activation is established, continued chemotherapy adds little and much toxicity, so it is withdrawn. The dual maintenance then sustains the immune response and vascular normalization indefinitely. Notably, none of the four patients completed the conventional six cycles of platinum-based chemotherapy, yet all derived meaningful benefit—a finding that challenges the rigid reliance on a fixed number of induction cycles and suggests that in early responders, fewer cycles may offer comparable efficacy with reduced cumulative harm.

Yet the report is candid about its limitations, and they are substantial. Baseline tumor tissue was never tested for PD-L1 expression, mismatch repair status, or tumor mutational burden, because the testing facilities were unavailable at the time of recurrence. The authors cannot exclude the possibility that one or more patients carried microsatellite instability–high disease, a biologically favorable subset in which checkpoint inhibitors work exceptionally well; if so, the impressive outcomes might reflect a lucky subset rather than broad utility of the MIMT strategy. Conversely, all four patients experienced benefit even though PD-1 monotherapy typically achieves low response rates in proficient mismatch repair tumors, raising the intriguing possibility that even microsatellite-stable patients may gain from this approach. The report also lacked immunological monitoring data—peripheral blood lymphocyte subsets, cytokine levels, or dynamic PD-L1 staining—leaving the biological basis of durable disease control unexplained from an immunological perspective.

Fistula safety deserves particular attention. Two patients had vesicovaginal and rectovaginal fistulas documented before starting bevacizumab-containing therapy, and in Case 2 the fistulas enlarged during treatment. The authors note that multiple confounding factors—local tumor progression directly invading the bladder and rectum, and prior pelvic radiotherapy including intensity-modulated radiotherapy and brachytherapy—are well-established risk factors for fistula development, so a causal relationship with bevacizumab cannot be definitively established, though a contributory role cannot be entirely excluded. They urge extra caution when administering bevacizumab to patients with fistulas. Similarly, the negative ascites cytology in Case 1, with false-negative rates as high as 30 to 37 percent reported in the literature, may have masked microscopic residual peritoneal disease, and the authors suggest that circulating tumor DNA or circulating tumor cells may eventually help predict outcomes more reliably than cytology alone.

What emerges from this case series is not a new standard of care but a provocative clinical signal wrapped in rigorous self-criticism. The authors explicitly state that their observations are intended to present experience with these cases and require validation in larger prospective studies. Given that recurrent or metastatic cervical cancer remains one of the most lethal gynecologic malignancies, and that prolonged triple-drug exposure until progression imposes considerable toxicity and financial burden, the hypothesis that a short induction burst followed by sustained dual maintenance might deliver comparable or superior survival with preserved quality of life is one that the field cannot afford to ignore. Whether the trigger-and-sustain pattern holds in randomized cohorts, and whether biomarkers such as PD-L1, mismatch repair status, and dynamic immune monitoring can identify which patients will respond, are questions that only larger prospective trials can answer. For now, four women in Yunnan have provided a glimpse of what might be possible when treatment timing is treated as a biological variable rather than a fixed protocol.

Subject of Research: Case Report: Prolonged survival with short-course triplet induction and dual maintenance in recurrent/metastatic cervical cancer

Article Title: Case Report: Prolonged survival with short-course triplet induction and dual maintenance in recurrent/metastatic cervical cancer

Article References: Duan, Y., Liu, D., Huang, Q., Man, J., Li, L., & Xue, Y. (2026). Case Report: Prolonged survival with short-course triplet induction and dual maintenance in recurrent/metastatic cervical cancer. Frontiers in Oncology, 16, Article 1961384. https://doi.org/10.3389/fonc.2026.1961384

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1961384

Keywords: Case, Report, Prolonged, survival, short-course, triplet, induction, dual, maintenance, recurrent, metastatic, cervical

The study, conducted by Dorcas Khosa of the Department of Criminology and Security Science at the University of South Africa, treats the cluster as an occasion for theory building rather than as a statistical sample. Using a structured comparison of four incidents selected against a stated threshold definition, Khosa asks what account of governance best explains the pattern of where the attacks happened, why they recurred, and how the state responded. The answer, she argues, lies not in any single institutional weakness but in the simultaneous attenuation of four kinds of authority at the same locations: regulatory, spatial, economic and intelligence. It is the conjunction of these attenuations, rather than any one of them, that produces the exposure on which coordinated attacks depend.

The venues themselves tell much of the story. Two of the four incidents occurred at unlicensed liquor outlets, premises the state has the statutory authority to license, inspect and close under the Gauteng Liquor Act of 2003 but does not effectively control. Police reported closing more than eleven thousand illegal taverns and making over eighteen thousand arrests for illegal liquor trading in a single six-month period, figures that describe a regime overwhelmed by an informal sector it cannot absorb rather than a successful enforcement campaign. South African research has long established an association between the density and trading hours of alcohol outlets and levels of interpersonal violence, making the regulatory gap a plausible contributor to exposure at these sites.

The firearms dimension compounds the failure. South Africa has a comparatively strict Firearms Control Act, yet the weapons recovered at several scenes, including AK-47 pattern rifles and nine-millimetre pistols, circulate in an illicit market the statute does not reach, replenished by theft, corruption in the chain of custody and cross-border trafficking. The case of a senior officer in the Gauteng firearm licence division who sold decommissioned police weapons into the gang economy over years, documented by crime analyst Mark Shaw, shows that the custody system inside the state itself can supply the illicit market. A regime that governs the licensed minority effectively while the illicit majority flows freely, Khosa argues, is regulating the wrong population.

Spatial governance forms the second domain of the diagnosis. The venues of the cluster, a hostel shebeen, a township tavern, an informal settlement and a taxi on a peri-urban road, all occupy the margins of formal urban administration. Informal settlements such as Jumpers lack formal addressing, reliable lighting, planned access routes and an effective municipal presence, making them partially illegible to the state, which cannot easily see, service or secure what it has not mapped. The assailants at Cleveland moved through the settlement for an hour in a space that afforded cover, multiple entrances and a low probability of timely response. Khosa is careful to note that legibility pursued as clearance would displace residents into equally illegible and equally exposed spaces, and she argues instead for upgrading in place, which serves both the resident’s claim on services and the state’s capacity to respond.

The third domain is economic. Bekkersdal and Cleveland both lie amid the abandoned gold workings of the Witwatersrand, where informal miners known as zama zamas operate in and around disused shafts. Research establishes that illegal mining in South Africa is organised, territorial and unusually violent by continental standards, diverting output estimated well in excess of fourteen billion rand a year and drawing migrant labour from Lesotho, Mozambique and Zimbabwe into its most exposed positions. At Meyerton both the victims and the suspects were Lesotho nationals, an observation about how the violence is embedded in transnational networks of labour, contraband and firearms rather than an argument about migration as a cause of crime, a claim the study explicitly examines and rejects. Where an extractive rent is available and lawful economic access to it is absent, territorial control passes to armed actors and disputes are settled by force.

The fourth domain concerns the security apparatus itself. The deficit here is specific: not a shortage of investigative competence, since suspects were identified and arrested in several cases, including eleven arrests within days at Bekkersdal, but the near-total absence of preventive intelligence. Coordinated attacks on known high-risk premises, in areas known to be contested by armed illicit economies, recurred without apparent interruption. Khosa formalises this as a reactive equilibrium: a security apparatus whose capability is concentrated after the event lowers the perceived probability of interruption, which lowers the expected cost of a coordinated attack and sustains recurrence. The signature of the attacks, assailants arriving and departing in vehicles and moving through a space in a sustained sweep, is consistent with an expectation of impunity.

Against this diagnosis, Khosa tests nine rival explanations, from under-policing and firearm availability to organised criminal competition, gang dynamics, prosecutorial attrition, state capture, migration, structural violence and contagion. Each accounts for part of the picture, she finds, but none explains siting, recurrence and the asymmetry between investigative success and preventive failure within a single account. The governance framework she develops, drawing on human security, nodal and polycentric governance and the security-development nexus, is organised around four principles: integration, prevention, inclusion and development. It is anchored in Sustainable Development Goal 16, which commits states to reduce violence and promote effective institutions, converting safety outcomes into development targets that are budgeted, tracked and publicly reported.

The framework’s central implication is its most provocative claim: accountability assigned by function will continue to leave the interstice ungoverned, because every authority can discharge its own mandate in full while the composite risk remains unowned. The institutional innovation the cluster demands, Khosa argues, is a governing node accountable for the space itself, such as a statutory Provincial Safety Governance Council with binding authority over police command, liquor licensing, municipalities, mining and home affairs functions. She proposes a sequenced programme: within twelve months, convening the council and auditing the police firearm chain of custody; within one to two years, statutory data-sharing obligations and a graduated liquor compliance pathway; over three to ten years, settlement upgrading at scale and regulation of small-scale mining.

Khosa is candid about the limits of her argument. The framework was built from four incidents in one province over twelve months, selected because the outcome occurred, and it has not been tested against comparable spaces where no attack followed. Its propositions are claims about a mechanism rather than findings about its strength, and she calls for a case-control study of incident and matched non-incident sites paired with interviews establishing what risk information existed before each attack. Yet the lesson she draws is specific enough to state without that work. The attacks occurred where four kinds of authority thinned out at once, and reactive intensification does not reach that condition, because force applied after the event operates on the consequences of interstitial exposure while leaving its production untouched. For unequal, rapidly urbanising societies across the global South, the decisive question is not how much security the state can supply but how well it can govern the margins where formal authority thins out. The alternative, she concludes, is to keep arriving after the shooting has stopped.

Subject of Research: Governance failures behind recurrent mass shootings in Gauteng, South Africa, and a proposed developmental security framework

Article Title: Recurrent mass shootings and the governance of insecurity in South Africa: towards a developmental security framework

Article References: Khosa, D. (2026). Recurrent mass shootings and the governance of insecurity in South Africa: towards a developmental security framework. Frontiers in Political Science, 8, Article 1929165. https://doi.org/10.3389/fpos.2026.1929165

Image Credits: AI Generated

DOI: 10.3389/fpos.2026.1929165

Keywords: South Africa, mass shootings, security governance, Gauteng, illegal mining, SDG 16, developmental security, crime prevention, Recurrent, mass, shootings, governance

The SLC29A3 gene encodes equilibrative nucleoside transporter 3, or ENT3, a protein essential for shuttling nucleosides across lysosomal and mitochondrial membranes. When biallelic mutations disrupt this transporter, the result is a multisystem disorder historically described under separate names, including H syndrome, Faisalabad histiocytosis, pigmented hypertrichosis with insulin-dependent diabetes, and familial Rosai-Dorfman disease. These entities are now recognized as a single clinical spectrum characterized by overlapping features such as histiocytic infiltration, lymphadenopathy, hyperpigmentation, hypertrichosis, and endocrine dysfunction. The clinical phenotype is remarkably diverse, with significant variability both between and within families, which often leads to delayed diagnosis, particularly in adults.

Born to first-degree cousin parents, the patient presented with a lifelong history of recurrent severe lower respiratory tract infections requiring multiple hospitalizations since childhood. She also suffered from chronic vomiting and recurrent nephrolithiasis, having previously undergone ureteroscopic lithotripsy and double-J stent placement. On physical examination, she exhibited significant growth failure, with height and weight below the 3rd percentile for her age. She had microcephaly, dysmorphic facial features, long and dense eyelashes, and generalized hirsutism. Skeletal findings included bilateral hallux valgus and mild flexion contractures of the fingers. Notably, her cognitive development was fully preserved.

Laboratory investigations revealed severe hypogammaglobulinemia, with IgG below 320 mg/dL against a reference range of 700 to 1600 mg/dL, IgA below 25 mg/dL against a reference of 70 to 400 mg/dL, and IgM below 25 mg/dL against a reference of 40 to 230 mg/dL. Hematologic evaluation showed microcytic anemia associated with iron deficiency. Liver function tests showed mildly elevated transaminases. Endocrine evaluation confirmed severe vitamin D deficiency with secondary hyperparathyroidism, although diabetes screening, including fasting blood glucose and HbA1c, was normal at the time of presentation. Chest computed tomography demonstrated chronic post-infectious small-airway changes, minimal bronchiectasis, and segmental consolidation in the right middle lobe. Abdominal imaging revealed right ureteric stones with associated hydronephrosis and a chronically small right kidney.

Genetic testing via whole-exome sequencing identified a homozygous missense variant in SLC29A3: c.1309G>A, resulting in the amino acid substitution p.Gly437Arg. The variant was classified as pathogenic according to American College of Medical Genetics and Genomics criteria and confirmed by Sanger sequencing. Parental segregation analysis was not performed because the parents were unavailable for testing. Copy-number variant analysis using an internally validated pipeline yielded no additional significant findings, establishing a diagnosis within the SLC29A3-related disorder spectrum.

The p.Gly437Arg variant identified in this patient is a known pathogenic allele frequently reported in Palestinian and Arab populations. In the original gene-discovery study, the variant was identified in several geographically clustered Arab families, and haplotype analysis showed a shared disease-associated haplotype among the analyzed families, consistent with a shared origin of the variant. Despite carrying a recurrent variant, this patient’s phenotype is notable for several reasons that expand the known clinical spectrum of the disease.

First, the severity of the immunologic phenotype is exceptional. While immunodeficiency and recurrent infections are recognized features of the SLC29A3-related disorder spectrum, the degree of hypogammaglobulinemia observed here, with nearly undetectable IgA and IgM and severely low IgG, is unusual. This raises the possibility that ENT3 deficiency may influence B-cell maturation or function, although the underlying mechanism remains unclear. Second, the associated renal manifestations, specifically recurrent nephrolithiasis and a small kidney, are noteworthy. Renal involvement has been reported in approximately 6 percent of cases within the spectrum, including hydronephrosis and renal asymmetry. It remains unclear whether these findings are a direct consequence of ENT3-related metabolic dysfunction or secondary to chronic inflammation and infection.

Following the diagnosis, a multidisciplinary management plan was initiated. Immunoglobulin replacement therapy was recommended and initiated, leading to a reduction in the frequency of respiratory infections during the initial six months of follow-up. The patient also received vitamin D supplementation and was referred to nephrology for management of recurrent stones and to endocrinology for ongoing diabetes surveillance. Comparison with previously reported cases of the same variant reveals significant phenotypic heterogeneity. While some patients present with predominant cutaneous and endocrine features, others exhibit a more severe immunologic and skeletal phenotype, suggesting the influence of genetic modifiers or environmental factors.

This case underscores the importance of considering the SLC29A3-related disorder spectrum in adults with longstanding multisystem disease. The combination of growth failure, skeletal anomalies, and immunodeficiency should prompt genetic evaluation even in the absence of the classic cutaneous hyperpigmentation often seen in H syndrome. Early genetic diagnosis using whole-exome sequencing is crucial for establishing a unifying diagnosis and guiding targeted multidisciplinary care, including immunoglobulin replacement therapy, which may substantially reduce infection-related morbidity, and vigilant surveillance for endocrine and renal complications. As a single case report, the study is limited by the absence of parental segregation analysis, functional studies of ENT3 activity, extended immune phenotyping, and detailed renal metabolic evaluation. Future studies are needed to determine whether nephrolithiasis is a primary metabolic manifestation of ENT3 deficiency or an incidental finding.

Subject of Research: Adult presentation of SLC29A3-related disorder spectrum with marked hypogammaglobulinemia due to a recurrent homozygous pathogenic variant (G437R): a case report

Article Title: Adult presentation of SLC29A3-related disorder spectrum with marked hypogammaglobulinemia due to a recurrent homozygous pathogenic variant (G437R): a case report

Article References: Ishqair, M. A., Masalma, A. K. H., Saleh, M., Awad, F., Khalili, A., Ishqair, A. A., Salah, S., & Saleh, A.-R. N. (2026). Adult presentation of SLC29A3-related disorder spectrum with marked hypogammaglobulinemia due to a recurrent homozygous pathogenic variant (G437R): a case report. Frontiers in Genetics, 17, Article 1904219. https://doi.org/10.3389/fgene.2026.1904219

Image Credits: AI Generated

DOI: 10.3389/fgene.2026.1904219

Keywords: Adult, presentation, SLC29A3-related, disorder, spectrum, marked, hypogammaglobulinemia, recurrent, homozygous, pathogenic, variant, G437R

The patient’s history is central to the case. Ten years earlier, he had undergone complete surgical excision of a basal cell carcinoma at another hospital with a 5-millimeter margin extending to the level of the subcutaneous tissue. Histopathological examination at that time had judged the surgical margins to be negative. Approximately one year before presentation at Kindai University Hospital, however, an erythematous plaque appeared at the same site and gradually enlarged. When the tumor was excised with what appeared to be adequate margins, the lesion was followed without further treatment, and it eventually recurred after a long interval.

At presentation, the clinical picture was deceptively simple. An ill-defined erythematous plaque measuring approximately 7 centimeters in diameter extended from the occipital region to the posterior neck, accompanied by a palpable subcutaneous nodule. Computed tomography showed a corresponding high-density lesion without lymph node enlargement or distant metastasis, and a skin biopsy confirmed recurrent basal cell carcinoma. The tumor was excised with a 1-centimeter margin, and the defect was reconstructed using a full-thickness skin graft.

Histopathological examination of the excised specimen revealed the technical crux of the case: infiltrative basal cell carcinoma with deep invasion and superficial lateral microextension extending beyond the clinically apparent lesion. The lateral margin was positive, and the minimum distance between the tumor and the deep surgical margin was only 0.5 millimeters. Infiltrative basal cell carcinoma is one of several histological subtypes in which tumor nests infiltrate in cord-like or micronodular patterns, which often obscure tumor boundaries both clinically and histopathologically.

Examination of the primary excision specimen showed that tumor nests were distributed predominantly within the subcutaneous tissue, with only limited continuity with the epidermis. Such a distribution may be difficult to recognize clinically as a superficial lesion and may lead to underestimation of tumor extent at the time of initial excision. Because the surgical margins had been judged negative, the lesion was followed without additional treatment, and the tumor eventually recurred after a long interval. The report suggests that this pattern of deep subcutaneous growth with limited epidermal continuity may explain why the tumor extent was difficult to appreciate clinically and why conventional margin determination based on surface findings was insufficient.

In the recurrent lesion, the tumor formed a nodular mass within the deeper subcutaneous tissue, suggesting a more aggressive growth pattern. At the same time, the tumor retained subtle lateral microextension within the superficial subcutaneous layer beyond the deep nodular component, making accurate margin determination difficult. Changes in skin coloration and a palpable subcutaneous nodule were used as clinical landmarks for estimating the tumor boundary; however, margin determination based solely on palpation was insufficient to capture the lateral tumor spread within the superficial subcutaneous layer.

To address this challenge, the surgical team employed a mapping biopsy strategy. Three weeks after the initial excision, mapping biopsies were performed at 13 sites to assess lateral tumor spread. The biopsies extended approximately 5 millimeters in depth to include the superficial subcutaneous tissue. Approximately 6 weeks later, additional excision was performed, with the resection boundary determined by connecting tumor-negative mapping biopsy sites. Histopathological examination confirmed complete tumor clearance, and no recurrence was observed at 6 months after the final excision.

Mohs micrographic surgery, a technique in which tumor is removed layer by layer with immediate microscopic examination of margins, is recommended for high-risk basal cell carcinoma. However, its availability varies among institutions. In retrospect, given the difficulty in clinically determining the tumor extent in this case, the authors suggest that temporary coverage with artificial dermis at the initial excision, followed by definitive reconstruction after histopathological confirmation of clear margins, might have been a reasonable alternative approach. In situations where Mohs surgery is not feasible, mapping biopsy may be useful for assessing tumor extent and guiding appropriate surgical margins.

The case underscores a broader principle in skin cancer surgery: histological subtype matters. In certain subtypes, including infiltrative basal cell carcinoma, tumor nests infiltrate in cord-like or micronodular patterns that often obscure tumor boundaries both clinically and histopathologically. It has also been reported that recurrent lesions may exhibit atypical patterns of tumor spread. The combination of deep nodular formation and superficial lateral microextension observed here represents one such atypical pattern, in which the clinically visible lesion substantially underestimated the true tumor extent in multiple directions.

For clinicians managing high-risk or recurrent basal cell carcinoma, the report offers a practical message. When tumor boundaries cannot be reliably determined clinically, particularly in infiltrative subtypes or recurrent lesions, surgeons should consider strategies that verify margins histopathologically before definitive reconstruction. Whether through Mohs micrographic surgery where available, temporary artificial dermis coverage pending margin confirmation, or mapping biopsy to delineate lateral spread, the goal remains the same: ensuring complete tumor clearance while avoiding unnecessary sacrifice of healthy tissue. In this case, a staged approach combining initial excision, systematic mapping biopsies, and definitive second excision achieved complete tumor clearance with no recurrence at 6 months.

Subject of Research: Recurrent infiltrative basal cell carcinoma with deep nodular formation and superficial lateral microextension

Article Title: Recurrent infiltrative basal cell carcinoma with deep nodular formation and superficial lateral microextension

Article References: Maeda, M., Yamamura, Y., Fujii, K., Nishimura, K., Usui, S., & Otsuka, A. (2026). Recurrent infiltrative basal cell carcinoma with deep nodular formation and superficial lateral microextension. Journal of Cutaneous Immunology and Allergy, 9, Article 17511. https://doi.org/10.3389/jcia.2026.17511

Image Credits: AI Generated

DOI: 10.3389/jcia.2026.17511

Keywords: basal cell carcinoma, skin cancer, tumor recurrence, mapping biopsy, surgical margins, infiltrative growth, Mohs surgery, dermatology, case report, subcutaneous invasion, Recurrent, infiltrative

The research addresses one of the most enduring puzzles in music cognition. Humans do not merely react to beats after they occur; we anticipate them, tapping slightly ahead of or behind the pulse in ways that follow lawful, measurable patterns. Decades of psychophysical experiments have catalogued these patterns in meticulous detail, including the negative mean asynchrony—the tendency for people to tap a few tens of milliseconds before the actual beat—and the way synchronization error grows predictably as tempo increases. What has remained contested is what kind of internal mechanism could give rise to these behaviors.

Traditional computational models have approached the problem in two main ways. Event-based models treat beat tracking as a discrete prediction problem, estimating the time of the next beat from a sequence of previous onsets. Dynamical systems models, by contrast, embed rhythm perception in continuous oscillatory processes, with neural oscillators that entrain to external periodic input. Each family of models captures part of the behavioral data, but each requires engineers to hand-design key structural assumptions about how timing information is represented. The new work asks a different question: if a general-purpose learning system is simply rewarded for staying in sync, will it discover these human-like dynamics on its own?

To find out, the researchers built recurrent neural networks—flexible computational systems whose internal activity depends on both their current input and their own prior states—and trained them using reinforcement learning. Rather than being given labeled examples of correct taps or supervised gradient signals tied to human behavior, the networks received a simple scalar reward for minimizing the error between their produced taps and the underlying beat of rhythmic sequences. This setup mirrors the ecological situation faced by a human learner, who must discover synchronization strategies through feedback rather than explicit instruction about where the beat lies.

After training on a diverse repertoire of rhythmic patterns and tempos, the networks were tested under the same conditions used in classic human tapping experiments. The results were notable on several fronts. The trained networks exhibited a negative mean asynchrony comparable in magnitude to that observed in human participants, and they reproduced the characteristic scaling of synchronization variability with tempo. When the rhythmic stimulus changed speed or switched pattern mid-sequence, the networks showed phase-correction and period-correction behaviors that closely matched human adaptation curves, including the asymmetric way people correct small errors more readily than large ones.

Perhaps most intriguingly, the internal dynamics of the trained networks revealed interpretable structure. Analysis of their recurrent activity showed convergence toward low-dimensional trajectories that effectively encoded beat phase and period, resembling the oscillator-like states posited by dynamical theories of beat perception. The networks had not been told to build internal clocks, yet the pressure of the reward function drove them to develop timing representations functionally equivalent to those long hypothesized in the neuroscience of rhythm. This suggests that human-like synchronization behavior may emerge from general learning principles rather than from rhythm-specific neural hardware alone.

The findings carry implications that extend beyond music cognition. Beat synchronization is increasingly used as a clinical and developmental marker: individual differences in tapping performance have been linked to language development, reading ability, and a range of neurological conditions, including Parkinson’s disease and developmental dyslexia. If a simple reinforcement-trained network can capture the canonical signatures of human performance, researchers gain a tractable model system for probing which aspects of synchronization depend on sensorimotor learning, on neural oscillation, or on the statistics of the rhythms themselves. Perturbing the artificial networks—lesioning units, altering reward structure, or changing input statistics—offers a level of experimental control impossible with human participants.

The work also speaks to a broader debate in artificial intelligence and neuroscience about the explanatory power of trained neural networks as scientific models. Critics argue that deep learning systems can fit behavior without illuminating mechanism. This study pushes back on that concern by showing that a behaviorally constrained network not only matched human outputs but also reproduced interpretable internal dynamics aligned with theoretical predictions. In this sense, the network functions as a computational hypothesis: it demonstrates that the reward structure of a synchronization task is sufficient, in principle, to generate the observed suite of human tapping phenomena, narrowing the space of explanations that cognitive science needs to consider.

Limitations remain, as the authors acknowledge. Real-world beat perception unfolds in acoustically rich environments, with polyphonic music, expressive timing, and social coordination among multiple players, whereas the trained networks operated on more controlled rhythmic stimuli. Human synchronization also engages motor physiology—the biomechanics of the limb, spinal circuitry, and cortical motor areas—that a simulated tapping output does not capture. Extending the approach to multimodal, embodied settings will be an important next step, as will testing whether networks trained on different reward schedules produce different asynchrony profiles, which could help explain the substantial individual variability seen in human tapping data.

Even with those caveats, the study marks a compelling convergence between machine learning and cognitive science. It suggests that the human sense of the beat—so effortless that we rarely notice it—may be the natural solution to a general predictive-control problem, one that an artificial learner rediscovers when faced with the same objective. As reinforcement-trained models continue to reproduce facets of perception, action, and now rhythmic coordination, the boundary between engineered and biological intelligence grows both more interesting and more informative to study.

Subject of Research: Reinforcement-trained recurrent neural networks as computational models of human beat synchronization and sensorimotor timing

Article Title: Reinforcement-trained recurrent networks reproduce human beat-synchronization dynamics

Article References: Ommi, Y., Yousefabadi, M., & Cannon, J. (2026). Reinforcement-trained recurrent networks reproduce human beat-synchronization dynamics. Communications Psychology. https://doi.org/10.1038/s44271-026-00529-1

Image Credits: AI Generated

DOI: 10.1038/s44271-026-00529-1

Keywords: beat synchronization, recurrent neural networks, reinforcement learning, music cognition, sensorimotor timing, rhythm perception, neural oscillators, tapping behavior, computational neuroscience, predictive timing, Reinforcement-trained, recurrent