diff --git a/src/skillspector/nodes/meta_analyzer.py b/src/skillspector/nodes/meta_analyzer.py
index 8f2b5410..87e478bd 100644
--- a/src/skillspector/nodes/meta_analyzer.py
+++ b/src/skillspector/nodes/meta_analyzer.py
@@ -282,6 +282,8 @@ def apply_filter(
         """
         _enrichment = tuple[str, str, float]
         confirmed_granular: dict[tuple[str, str, int, int | None], _enrichment] = {}
+        # Fallback index keyed without end_line (see lookup below). Issue #67.
+        confirmed_by_start: dict[tuple[str, str, int], _enrichment] = {}
         confirmed_coarse: dict[tuple[str, str], _enrichment] = {}
 
         for batch, llm_items in batch_results:
@@ -308,6 +310,7 @@ def apply_filter(
                             int(end_line) if end_line is not None else None,
                         )
                     ] = enrichment
+                    confirmed_by_start[(file_path, pattern_id, int(start_line))] = enrichment
                 else:
                     confirmed_coarse[(file_path, pattern_id)] = enrichment
 
@@ -316,10 +319,13 @@ def apply_filter(
             exact_key = (f.file, f.rule_id, f.start_line, f.end_line)
             start_only_key = (f.file, f.rule_id, f.start_line, None)
             coarse_key = (f.file, f.rule_id)
+            start_key = (f.file, f.rule_id, f.start_line) if f.start_line is not None else None
             if exact_key in confirmed_granular:
                 expl, rem, conf = confirmed_granular[exact_key]
             elif start_only_key in confirmed_granular:
                 expl, rem, conf = confirmed_granular[start_only_key]
+            elif f.end_line is None and start_key is not None and start_key in confirmed_by_start:
+                expl, rem, conf = confirmed_by_start[start_key]
             elif coarse_key in confirmed_coarse:
                 expl, rem, conf = confirmed_coarse[coarse_key]
             else:
diff --git a/tests/nodes/test_meta_analyzer.py b/tests/nodes/test_meta_analyzer.py
new file mode 100644
index 00000000..0087f463
--- /dev/null
+++ b/tests/nodes/test_meta_analyzer.py
@@ -0,0 +1,100 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Unit tests for LLMMetaAnalyzer.apply_filter (no LLM / no network)."""
+
+from skillspector.llm_analyzer_base import Batch
+from skillspector.models import Finding
+from skillspector.nodes.meta_analyzer import LLMMetaAnalyzer
+
+
+def _analyzer() -> LLMMetaAnalyzer:
+    # Skip __init__ so no LLM client / API key is needed; apply_filter is pure.
+    return LLMMetaAnalyzer.__new__(LLMMetaAnalyzer)
+
+
+def _finding(rule_id: str, start_line: int, end_line: int | None = None) -> Finding:
+    return Finding(
+        rule_id=rule_id,
+        message=f"static finding {rule_id}",
+        severity="CRITICAL",
+        confidence=0.9,
+        file="requirements.txt",
+        start_line=start_line,
+        end_line=end_line,
+    )
+
+
+def _llm_item(rule_id: str, start_line: int, **kw: object) -> dict[str, object]:
+    item: dict[str, object] = {
+        "pattern_id": rule_id,
+        "is_vulnerability": True,
+        "confidence": 1.0,
+        "start_line": start_line,
+        "_file": "requirements.txt",
+    }
+    item.update(kw)
+    return item
+
+
+def test_confirmed_finding_kept_when_model_returns_end_line() -> None:
+    """Regression: a static finding with end_line=None must still match a
+    confirmation whose end_line is populated (e.g. end_line == start_line, as
+    some models return). Previously these confirmed findings were silently
+    dropped. See issue #67."""
+    findings = [_finding("SC4", 4), _finding("SC4", 5)]
+    items = [_llm_item("SC4", 4, end_line=4), _llm_item("SC4", 5, end_line=5)]
+    batch = Batch(file_path="requirements.txt", content="", findings=findings)
+
+    kept = _analyzer().apply_filter(findings, [(batch, items)])
+
+    assert {f.start_line for f in kept} == {4, 5}
+    assert len(kept) == 2
+
+
+def test_rejected_finding_still_dropped() -> None:
+    """The end_line-agnostic fallback must not resurrect findings the LLM
+    rejected (is_vulnerability=False)."""
+    findings = [_finding("SC4", 4)]
+    items = [_llm_item("SC4", 4, end_line=4, is_vulnerability=False)]
+    batch = Batch(file_path="requirements.txt", content="", findings=findings)
+
+    kept = _analyzer().apply_filter(findings, [(batch, items)])
+
+    assert kept == []
+
+
+def test_low_confidence_finding_dropped() -> None:
+    """Confirmations below the confidence threshold are not kept."""
+    findings = [_finding("SC4", 4)]
+    items = [_llm_item("SC4", 4, end_line=4, confidence=0.3)]
+    batch = Batch(file_path="requirements.txt", content="", findings=findings)
+
+    kept = _analyzer().apply_filter(findings, [(batch, items)])
+
+    assert kept == []
+
+
+def test_exact_end_line_match_still_works() -> None:
+    """Existing behaviour: when both sides carry the same concrete end_line,
+    the finding is kept (no regression from the new fallback)."""
+    findings = [_finding("AST1", 21, end_line=21)]
+    items = [_llm_item("AST1", 21, end_line=21)]
+    batch = Batch(file_path="requirements.txt", content="", findings=findings)
+
+    kept = _analyzer().apply_filter(findings, [(batch, items)])
+
+    assert len(kept) == 1
+    assert kept[0].rule_id == "AST1"